ArXiv Daily - AI for Materials Science!

Inverse design aims to find design parameters $x$ achieving target performance $y^*$. Generative approaches learn bidirectional mappings between designs and labels, enabling diverse solution sampling. However, standard conditional flow matching (CFM), when adapted to inverse problems by pairing labels with design parameters, exhibits strong sensitivity to their arbitrary ordering and scaling, leading to unstable training. We introduce Diagonal Flow Matching (Diag-CFM), which resolves this through a zero-anchoring strategy that pairs design coordinates with noise and labels with zero, making the learning problem provably invariant to coordinate permutations. This yields order-of-magnitude improvements in round-trip accuracy over CFM and invertible neural network baselines across design dimensions up to $P{=}100$. We develop two architecture-intrinsic uncertainty metrics, Zero-Deviation and Self-Consistency, that enable three practical capabilities: selecting the best candidate among multiple generations, abstaining from unreliable predictions, and detecting out-of-distribution targets; consistently outperforming ensemble and general-purpose alternatives across all tasks. We validate on airfoil, gas turbine combustor, and an analytical benchmark with scalable design dimension.

CO2 reduction requires efficient catalysts, yet materials discovery remains bottlenecked by 10-20 year development cycles requiring deep domain expertise. This paper demonstrates how large language models can assist the catalyst discovery process by helping researchers explore chemical spaces and interpret results when augmented with retrieval-based grounding. We introduce a retrieval-augmented generation framework that enables GPT-4 to navigate chemical space by accessing a database of 50,000+ known materials, adapting general-purpose language understanding for high-throughput materials design. Our approach generated over 250 catalyst candidates with an 82% thermodynamic stability rate while addressing multi-objective constraints: 68% achieved <$100/kg cost with metallic conductivity (band gap<0.1eV) and mechanical stability (B/G>1.75). The best-performing Fe0.2Co0.2Ni0.2Ir0.1Ru0.3 achieves 0.285V limiting potential (25% improvement over IrO2), while Cr0.2Fe0.2Co0.3Ni0.2Mo0.1 optimally balances performance-cost trade-offs at $18/kg. Volcano plot analysis confirms that 78% of LLM-generated catalysts cluster near the theoretical activity optimum, while our system achieves 200x computational efficiency compared to traditional high-throughput screening. By demonstrating that retrieval-augmented generation can ground AI creativity in physical constraints without sacrificing exploration, this work demonstrates an approach where natural language interfaces can streamline materials discovery workflows, enabling researchers to explore chemical spaces more efficiently while the LLM assists in result interpretation and hypothesis generation.

Cold metals are a class of metals with an intrinsic energy gap located close to the Fermi level, which enables cold-carrier injection for steep-slope transistors and is therefore promising for low-power electronic applications. High-throughput screening has revealed 252 three-dimensional (3D) cold metals in the Materials Project database, but database searches are inherently limited to known compounds. Here we present an inverse-design workflow that generates 3D cold metals using MatterGPT, a conditional autoregressive Transformer trained on SLICES, an invertible and symmetry-invariant crystal string representation. We curate a training set of 26,309 metallic structures labeled with energy above hull and a unified band-edge distance descriptor that merges p-type and n-type cold-metal characteristics to address severe label imbalance. Property-conditioned generation targeting thermodynamic stability and 50-500 meV band-edge distances produces 148,506 unique candidates; 92.1% are successfully reconstructed to 3D structures and down-selected by symmetry, uniqueness and novelty filters, followed by high-throughput DFT validation. We identify 257 cold metals verified as novel with respect to the Materials Project database, with gaps around the Fermi level spanning 50-500 meV. First-principles phonon, electronic-structure, and work-function calculations for representative candidates confirm dynamical stability and contact-relevant work functions. Our results demonstrate that SLICES-enabled generative transformers can expand the chemical space of cold metals beyond high-throughput screening, providing a route to low-power electronic materials discovery.

Many of the most important problems in science and engineering are inverse problems: given a desired outcome, find a design that achieves it. Evaluating whether a candidate meets the spec is often routine; a binding energy can be computed, a reactor yield simulated, a pharmacokinetic profile predicted. But searching a combinatorial design space for inputs that satisfy those targets is fundamentally harder. We introduce SciDesignBench, a benchmark of 520 simulator-grounded tasks across 14 scientific domains and five settings spanning single-shot design, short-horizon feedback, long-horizon refinement, and seed-design optimization. On the 10-domain shared-core subset, the best zero-shot model reaches only 29.0% success despite substantially higher parse rates. Simulator feedback helps, but the leaderboard changes with horizon: Sonnet 4.5 is strongest in one-turn de novo design, whereas Opus 4.6 is strongest after 20 turns of simulator-grounded refinement. Providing a starting seed design reshuffles the leaderboard again, demonstrating that constrained modification requires a fundamentally different capability from unconstrained de novo generation. We then introduce RLSF, a simulator-feedback training recipe. An RLSF-tuned 8B model raises single-turn success rates by 8-17 percentage points across three domains. Together, these results position simulator-grounded inverse design as both a benchmark for scientific reasoning and a practical substrate for amortizing expensive test-time compute into model weights.

Today's scientific challenges, from climate modeling to Inertial Confinement Fusion design to novel material design, require exploring huge design spaces. In order to enable high-impact scientific discovery, we need to scale up our ability to test hypotheses, generate results, and learn from them rapidly. We present MADA (Multi-Agent Design Assistant), a Large Language Model (LLM) powered multi-agent framework that coordinates specialized agents for complex design workflows. A Job Management Agent (JMA) launches and manages ensemble simulations on HPC systems, a Geometry Agent (GA) generates meshes, and an Inverse Design Agent (IDA) proposes new designs informed by simulation outcomes. While general purpose, we focus development and validation on Richtmyer--Meshkov Instability (RMI) suppression, a critical challenge in Inertial Confinement Fusion. We evaluate on two complementary settings: running a hydrodynamics simulations on HPC systems, and using a pre-trained machine learning surrogate for rapid design exploration. Our results demonstrate that the MADA system successfully executes iterative design refinement, automatically improving designs toward optimal RMI suppression with minimal manual intervention. Our framework reduces cumbersome manual workflow setup, and enables automated design exploration at scale. More broadly, it demonstrates a reusable pattern for coupling reasoning, simulation, specialized tools, and coordinated workflows to accelerate scientific discovery.

Corrosion testing is slow, labor-intensive, and sensitive to operator technique, limiting the generation of large, high-quality datasets for data-driven materials discovery. We introduce the Materials Acceleration Platform for Electrochemistry (MAP-E), an autonomous, high-throughput system capable of performing parallel electrochemical experiments. MAP-E integrates robotic liquid handling, sample transfer, and multi-channel potentiostatic control and extract corrosion metrics without human intervention. Validation against an ASTM G61-analog benchmark demonstrates reproducibility, with a standard deviation of 76 mV in pitting potential across 32 automated measurements. The platform was then employed to autonomously construct pH-chloride stability diagrams for 304 stainless steel using an uncertainty-driven sampling strategy on a Gaussian Process surrogate model. This approach reduces operator involvement and accelerates the exploration of environmental spaces. MAP-E establishes a framework for autonomous electrochemical experimentation, enabling generation of corrosion datasets that inform materials discovery, alloy design, and durability assessment in service environments.

Bridging generative foundation models with non-equilibrium thin-film synthesis remains a central challenge, limiting the practical impact of AI-driven materials discovery on semiconductor dielectrics. Here, we introduce IDEAL (Inverse Design for Experimental Atomic Layers), an inverse-design platform that links generative diffusion models, machine learning interatomic potentials, and graph neural network property predictors with atomic layer deposition (ALD). We demonstrate IDEAL using the Hf-Zr-O system as a stringent benchmark for semiconductor-relevant complex oxides. The platform statistically enumerates thermodynamically plausible structures and constructs a composition-structure-property map. Crucially, it identifies a narrow composition window where low-energy tetragonal and orthorhombic phases cluster, revealing trade-offs between band gap and dielectric response. Experimental validation using atomic layer modulation (ALM) corroborates these predictions, demonstrating predictive guidance under realistic, non-equilibrium thin-film growth. By experimentally closing the loop, IDEAL provides a transferable and generalizable route to the precision synthesis of next-generation semiconductor dielectrics.

Recent advances in deep learning inspired neural network-based approaches to computational materials discovery (CMD). A plethora of problems in this field involve finding materials that optimize a target property. Nevertheless, the increasingly popular generative modeling methods are ineffective at boldly exploring attractive regions of the materials space due to their maximum likelihood training. In this work, we offer an alternative CMD technique based on offline model-based optimization (MBO) that fuses direct optimization of a target material property into generation. To that end, we introduce a domain-specific model, dubbed CliqueFlowmer, that incorporates recent advances of clique-based MBO into transformer and flow generation. We validate CliqueFlowmer's optimization abilities and show that materials it produces strongly outperform those provided by generative baselines. To enable employment of CliqueFlowmer in specialized materials optimization problems and support interdisciplinary research, we open-source our code at https://github.com/znowu/CliqueFlowmer.

Recent advances in deep learning inspired neural network-based approaches to computational materials discovery (CMD). A plethora of problems in this field involve finding materials that optimize a target property. Nevertheless, the increasingly popular generative modeling methods are ineffective at boldly exploring attractive regions of the materials space due to their maximum likelihood training. In this work, we offer an alternative CMD technique based on offline model-based optimization (MBO) that fuses direct optimization of a target material property into generation. To that end, we introduce a domain-specific model, dubbed CliqueFlowmer, that incorporates recent advances of clique-based MBO into transformer and flow generation. We validate CliqueFlowmer's optimization abilities and show that materials it produces strongly outperform those provided by generative baselines. To enable employment of CliqueFlowmer in specialized materials optimization problems and support interdisciplinary research, we open-source our code at https://github.com/znowu/CliqueFlowmer.

Crystal structures define how matter is organized at the atomic level. In the realm of crystalline inorganic materials, new structure types are rarely found, and most experimentally-realized structural motifs were established decades ago. Considerable efforts are underway to discover new crystalline inorganic compounds, often aided by artificial intelligence (AI). However, thus far, these methods have not yielded convincing new structure types, but rather substitutional variations of existing compounds. Here we introduce a new structure type adopted by the compound GdNiSn4, discovered the old-fashioned way. We test whether current state-of-the-art AI-based material generation models can predict this material in its correct structure and find that they fail to do so. We carefully analyze the new structure and argue that it can be viewed as a stack of two known structure types. We explore electronic and steric factors underlying its stability and propose strategies to improve future AI-guided materials discovery. Furthermore, we report complex magnetic properties in GdNiSn4, highlighting its potential interest for future studies of unconventional magnetism.

Magnonic crystals (MCs) are emerging spintronic metamaterials capable of manipulating transmission properties of magnons, the quanta of spin waves. Due to the complex relationship between lattice geometry and magnonic band dispersion, it remains challenging to establish general design strategies for optimizing targeted properties in MCs. In this study, we demonstrated an inverse-design framework for two-dimensional MCs to explore unconventional lattice structures with large magnonic band gaps. We employed genetic algorithms to enable global exploration of structures with a complete band gap as the objective property, and used frequency-domain micromagnetic simulations for computationally efficient band gap evaluation. Our established inverse-design method successfully discovered several previously unreported designs of MCs, whose performance was validated using time-domain micromagnetic simulations. Furthermore, we observed that the design landscape becomes increasingly non-convex at high-order bands, suggesting the existence of multiple design solutions. The overall inverse-design framework is expected to be broadly applicable to experimentally accessible material systems and device dimensions, facilitating the formulation of design rules for MCs.

Generative models hold great promise for accelerating material discovery but are often limited by their inflexible single-stage generative process in designing valid and diverse materials. To address this, we propose a two-stage generative framework, Lang2Str, that combines the strengths of large language models (LLMs) and flow-based models for flexible and precise material generation. Our method frames the generative process as a conditional generative task, where an LLM provides high-level conditions by generating descriptions of material unit cells' geometric layouts and properties. These descriptions, informed by the LLM's extensive background knowledge, ensure reasonable structure designs. A conditioned flow model then decodes these textual conditions into precise continuous coordinates and unit cell parameters. This staged approach combines the structured reasoning of LLMs and the distribution modeling capabilities of flow models. Experimental results show that our method achieves competitive performance on \textit{ab initio} material generation and crystal structure prediction tasks, with generated structures exhibiting closer alignment to ground truth in both geometry and energy levels, surpassing state-of-the-art models. The flexibility and modularity of our framework further enable fine-grained control over the generation process, potentially leading to more efficient and customizable material design.

Stochastic differential equation (SDE)-based generative models have achieved substantial progress in conditional generation via training-free differentiable loss-guided approaches. However, existing methodologies utilizing posterior sam- pling typically confront a substantial estimation error, which results in inaccu- rate gradients for guidance and leading to inconsistent generation results. To mitigate this issue, we propose that performing an additional backward denois- ing step and Monte-Carlo sampling (ABMS) can achieve better guided diffu- sion, which is a plug-and-play adjustment strategy. To verify the effectiveness of our method, we provide theoretical analysis and propose the adoption of a dual-focus evaluation framework, which further serves to highlight the critical problem of cross-condition interference prevalent in existing approaches. We conduct experiments across various task settings and data types, mainly includ- ing conditional online handwritten trajectory generation, image inverse problems (inpainting, super resolution and gaussian deblurring) molecular inverse design and so on. Experimental results demonstrate that our approach can be effec- tively used with higher order samplers and consistently improves the quality of generation samples across all the different scenarios.

We introduce Large Electron Model, a single neural network model that produces variational wavefunctions of interacting electrons over the entire Hamiltonian parameter manifold. Our model employs the Fermi Sets architecture, a universal representation of many-body fermionic wavefunctions, which is further conditioned on Hamiltonian parameter and particle number. On interacting electrons in a two-dimensional harmonic potential, a single trained model accurately predicts the ground state wavefunction while generalizing across unseen coupling strengths and particle-number sectors, producing both accurate real-space charge densities and ground state energies, even up to $50$ particles. Our results establish a foundation model method for material discovery that is grounded in the variational principle, while accurately treating strong electron correlation beyond the capacity of density functional theory.

Nonbenzenoid carbon frameworks expand low-dimensional material design via controlled asymmetry. Here, we show the experimentally realized 4-5-6-8 carbon nanoribbon establishes a topology-driven paradigm for multiproperty engineering, not just a graphene variant. Using hybrid DFT, tight-binding, and molecular dynamics in a multiscale framework, we demonstrate the symmetry-broken lattice stabilizes hierarchical bonds within standard energy ranges. This geometry produces a robust semiconducting state (hybrid gap >1 eV) and enables strain as a controllable modulation parameter. A tight-binding Hamiltonian fitted only at equilibrium accurately captures strain-dependent band evolution, proving the essential physics is topology-dominated. Mechanical analysis reveals high stiffness with fracture governed by the largest polygons, showing asymmetry redistributes stress without compromising integrity. Intrinsic phonon scattering suppresses thermal conductance, enabling favorable thermoelectric performance without extrinsic disorder. Optical response confirms non-equivalent ring connectivity reorganizes interband transitions, promoting strong visible absorption and efficient photocarrier generation. These results position topology as a governing parameter coupling elasticity, electronics, thermal transport, and optics, establishing the 4-5-6-8 nanoribbon as a unified platform for predictive design of multifunctional carbon materials.

Modern materials science has historically been founded on combining restricted subsets of the periodic table, favoring high-purity, few-element systems. However, the demands of an emerging circular economy, together with the need to understand materials behavior under planetary and industrial extremes, increasingly require mastering Mendeleev materials - chemically and structurally complex systems that span large portions of the periodic table. In these regimes, current universal machine-learning interatomic potentials often fail, largely due to systematic gaps in traditional training datasets that heavily emphasize low-energy, near-equilibrium structures. We address this limitation by introducing a chemistry-agnostic, information-entropy-maximization protocol for data generation. By decoupling structural sampling from thermodynamic bias, our approach provides a robust physical prior for atomic interactions across the entire periodic table, including regimes far from equilibrium and under extreme conditions. Training a Graph Atomic Cluster Expansion (GRACE) model on the resulting statistically maximized entropy (SMAX) dataset yields markedly improved robustness across a range of stringent benchmarks. These include large-strain phase transformations in tin, defect evolution in tungsten-based alloys, and catalytic reaction barrier prediction. More broadly, our approach establishes a scalable and principled methodology for navigating the vast chemical and configurational space relevant to future materials design. It enables a paradigm of discovery by simulation in which unbiased sampling protocols autonomously resolve emergent structures in multi-elemental mixtures-such as systems containing the nine most abundant elements in the Earth's crust-without reliance on a priori chemical assumptions.

2D materials exhibiting in-plane anisotropy enable novel functionality in electronic, optoelectronic, and photonic devices, yet their availability is generally limited to naturally-occurring low-symmetry van der Waals compounds. Here, we demonstrate an approach to structural engineering in a family of blue-emitting 2D silver phenylchalcogenide semiconductors based on steric interactions among surface-bound organic molecular ligands. By strategically halogenating specific sites of phenyl ligands, we demonstrate dramatic changes to the inorganic AgSe plane in mithrene (silver phenylselenolate, AgSePh). Density functional theory revealed pronounced in-plane electronic anisotropy for direct-gap fluorinated derivatives, while a chlorinated variant exhibited a direct-to-indirect bandgap transition. Furthermore, some fluorinated variants displayed strongly polarized absorption and luminescence, accompanied by a 10x enhancement in photoluminescence quantum yield. This work establishes a versatile approach for tailoring optoelectronic properties in hybrid semiconductors that is difficult or impossible to achieve in all-inorganic materials alone, offering new opportunities in advanced material design.

The discovery of novel carbon allotropes with tailored thermal and mechanical properties is critical for advanced thermal management. However, exploring the vast configurational space of carbon using \textit{ab initio} calculations remains computationally prohibitive. Driven by the rich topological landscape of carbon, where the competition between $sp, sp^2,$ and $sp^3$ hybridization states dictates material performance, we establish a closed-loop AI framework to explore this complex configurational space. We introduce a hybridization entropy descriptor to guide the search beyond conventional forms. Here, we establish a closed-loop AI framework that synergizes a Large Language Model (LLM) for structural generation with a Machine Learning Potential (MLP) for accelerated evaluation. Leveraging CrystaLLM to generate candidates and an iteratively refined MLP for high-fidelity validation, we screened thousands of structures to identify several stable allotropes with exotic properties. Specifically, we report ``yne-diamond C$_{12}$'' and ``yne-hex-diamond C$_{8}$'', which exhibit extreme thermal anisotropy and ultralow in-plane shear stiffness arising from their mixed $sp$-$sp^3$ hybridization. Furthermore, we discovered a complex $sp$-$sp^2$-$sp^3$ hybridized C$_{12}$ phase that combines metallic conductivity with an anomalous negative Poisson's ratio. Notably, we identified a superhard phase (C16_3) possessing a calculated Vickers hardness (103.3 GPa) exceeding that of diamond 96 GPa). Microscopic analysis reveals that thermal transport in these materials is governed by the interplay between rigid frameworks and flexible linkers. This work expands the known carbon phase space and demonstrates the efficacy of coupling generative AI with machine learning potentials for the accelerated inverse design of functional materials.

Entropy governs molecular self-assembly, phase transitions, and material stability, yet remains challenging to quantify and directly control in molecular systems. Here, we demonstrate that the computable information density (CID), a data compression-based information theoretic metric, provides an instantaneous general measure of configurational entropy in molecular dynamics simulations, reflecting both local and long-range structural organization. We validate the CID across systems of increasing complexity, beginning with single-component Lennard-Jones melting before examining binary phase separation, polymer condensation and dispersion, and assembly of amorphous carbon networks at multiple densities. Unlike conventional order parameters, CID requires no a priori knowledge of relevant structural features and captures entropic signatures across a variety of molecular systems and discretization resolutions. By establishing entropy as a directly accessible structural metric, this framework lays a foundation for future entropy-driven materials design and optimization strategies.

Large language models (LLMs) are becoming increasingly applied beyond natural language processing, demonstrating strong capabilities in complex scientific tasks that traditionally require human expertise. This progress has extended into materials discovery, where LLMs introduce a new paradigm by leveraging reasoning and in-context learning, capabilities absent from conventional machine learning approaches. Here, we present a Multi-Agent-based Electrocatalyst Search Through Reasoning and Optimization (MAESTRO) framework in which multiple LLMs with specialized roles collaboratively discover high-performance single atom catalysts for the oxygen reduction reaction. Within an autonomous design loop, agents iteratively reason, propose modifications, reflect on results and accumulate design history. Through in-context learning enabled by this iterative process, MAESTRO identified design principles not explicitly encoded in the LLMs' background knowledge and successfully discovered catalysts that break conventional scaling relations between reaction intermediates. These results highlight the potential of multi-agent LLM frameworks as a powerful strategy to generate chemical insight and discover promising catalysts.

Exact stochastic simulation of continuous-time Markov chains (CTMCs) is essential when discreteness and noise drive system behavior, but the hard categorical event selection in Gillespie-type algorithms blocks gradient-based learning. We eliminate this constraint by decoupling forward simulation from backward differentiation, with hard categorical sampling generating exact trajectories and gradients propagating through a continuous massively-parallel Gumbel-Softmax straight-through surrogate. Our approach enables accurate optimization at parameter scales over four orders of magnitude beyond existing simulators. We validate for accuracy, scalability, and reliability on a reversible dimerization model (0.09% error), a genetic oscillator (1.2% error), a 203,796-parameter gene regulatory network achieving 98.4% MNIST accuracy (a prototypical deep-learning multilayer perceptron benchmark), and experimental patch-clamp recordings of ion channel gating (R^2 = 0.987) in the single-channel regime. Our GPU implementation delivers 1.9 billion steps per second, matching the scale of non-differentiable simulators. By making exact stochastic simulation massively parallel and autodiff-compatible, our results enable high-dimensional parameter inference and inverse design across systems biology, chemical kinetics, physics, and related CTMC-governed domains.

Quantum chemistry calculations are a key component of the materials discovery process. The results from first-principles explorations enable the prediction of material properties prior to experimental validation. Despite their impact, the practical use of first-principles methods remains limited by the expertise required to design, execute, and troubleshoot complex computational workflows. Even when workflows are successfully built, they are sometimes rigid and not adaptable to different use cases. Recent advances in large language models (LLMs) and agentic systems offer a pathway to flexibly automate these processes and lower barriers to entry. Here, we introduce El Agente Sólido, a hierarchical multi-agent framework for automating solid-state quantum chemistry workflows using the open-source Quantum ESPRESSO simulation package. The framework translates high-level scientific objectives expressed in natural language into end-to-end computational pipelines that include structure generation, input file construction, workflow execution, and post-processing analysis. El Agente Sólido integrates density functional theory with phonon calculations and machine-learning interatomic potentials to enable efficient and physically consistent simulations. Extensive benchmarking and case studies demonstrate that El Agente Sólido reliably executes a wide range of solid-state calculations, highlighting its potential to improve reproducibility and accelerate computational materials discovery

Crystal structure prediction (CSP), which aims to predict the three-dimensional atomic arrangement of a crystal from its composition, is central to materials discovery and mechanistic understanding. However, given the composition and atomic counts in a unit cell, existing methods struggle with the NP-hard combinatorial challenge of rigorous symmetry enforcement or rely on retrieving known templates, which inherently limits both physical fidelity and the ability to discover genuinely new materials. To solve this, we propose a symmetry-driven generative framework. Our approach leverages large language models to encode chemical semantics and directly generate fine-grained Wyckoff patterns from atomic stoichiometry and counts, effectively circumventing the limitations inherent to database lookups. Crucially, to overcome the exponentially complex problem of combinatorial site assignments, we incorporate domain knowledge through an efficient, linear-complexity heuristic beam search algorithm that rigorously enforces algebraic consistency between site multiplicities and atomic stoichiometry and counts. By integrating this symmetry-consistent template into a diffusion backbone, our approach constrains the stochastic generative trajectory to a physically valid geometric manifold. This framework achieves state-of-the-art performance across stability, uniqueness, and novelty (SUN) benchmarks, alongside superior matching performance, thereby establishing a new paradigm for the rigorous exploration of targeted crystallographic space which can be previously uncharted, with no reliance on existing databases or a priori structural knowledge.

Crystal structure prediction (CSP), which aims to predict the three-dimensional atomic arrangement of a crystal from its composition, is central to materials discovery and mechanistic understanding. Existing deep learning models often treat crystallographic symmetry only as a soft heuristic or rely on space group and Wyckoff templates retrieved from known structures, which limits both physical fidelity and the ability to discover genuinely new material structures. In contrast to retrieval-based methods, our approach leverages large language models to encode chemical semantics and directly generate fine-grained Wyckoff patterns from composition, effectively circumventing the limitations inherent to database lookups. Crucially, we incorporate domain knowledge into the generative process through an efficient constrained-optimization search that rigorously enforces algebraic consistency between site multiplicities and atomic stoichiometry. By integrating this symmetry-consistent template into a diffusion backbone, our approach constrains the stochastic generative trajectory to a physically valid geometric manifold. This framework achieves state-of-the-art performance across stability, uniqueness, and novelty (SUN) benchmarks, alongside superior matching performance, thereby establishing a new paradigm for the rigorous exploration of targeted crystallographic space. This framework enables efficient expansion into previously uncharted materials space, eliminating reliance on existing databases or a priori structural knowledge.

Crystal structure prediction (CSP), which aims to predict the three-dimensional atomic arrangement of a crystal from its composition, is central to materials discovery and mechanistic understanding. However, given the composition in a unit cell, existing methods struggle with the NP-hard combinatorial challenge of rigorous symmetry enforcement or rely on retrieving known templates, which inherently limits both physical fidelity and the ability to discover genuinely new materials. To solve this, we propose a symmetry-driven generative framework. Our approach leverages large language models to encode chemical semantics and directly generate fine-grained Wyckoff patterns from atomic stoichiometry, effectively circumventing the limitations inherent to database lookups. Crucially, to overcome the exponentially complex problem of combinatorial site assignments, we incorporate domain knowledge through an efficient, linear-complexity heuristic beam search algorithm that rigorously enforces algebraic consistency between site multiplicities and atomic stoichiometry. By integrating this symmetry-consistent template into a diffusion backbone, our approach constrains the stochastic generative trajectory to a physically valid geometric manifold. This framework achieves state-of-the-art performance across stability, uniqueness, and novelty (SUN) benchmarks, alongside superior matching performance, thereby establishing a new paradigm for the rigorous exploration of targeted crystallographic space which can be previously uncharted, with no reliance on a priori structural knowledge.

The inverse design of RNA three-dimensional (3D) structures is crucial for engineering functional RNAs in synthetic biology and therapeutics. While recent deep learning approaches have advanced this field, they are typically optimized and evaluated using native sequence recovery, which is a limited surrogate for structural fidelity, since different sequences can fold into similar 3D structures and high recovery does not necessarily indicate correct folding. To address this limitation, we propose RIDER, an RNA Inverse DEsign framework with Reinforcement learning that directly optimizes for 3D structural similarity. First, we develop and pre-train a GNN-based generative diffusion model conditioned on the target 3D structure, achieving a 9% improvement in native sequence recovery over state-of-the-art methods. Then, we fine-tune the model with an improved policy gradient algorithm using four task-specific reward functions based on 3D self-consistency metrics. Experimental results show that RIDER improves structural similarity by over 100% across all metrics and discovers designs that are distinct from native sequences.

Inverse design problems are common in engineering and materials science. The forward direction, i.e., computing output quantities from design parameters, typically requires running a numerical simulation, such as a FEM, as an intermediate step, which is an optimization problem by itself. In many scenarios, several design parameters can lead to the same or similar output values. For such cases, multi-modal probabilistic approaches are advantageous to obtain diverse solutions. A major difficulty in inverse design stems from the structure of the design space, since discrete parameters or further constraints disallow the direct use of gradient-based optimization. To tackle this problem, we propose a novel inverse design method based on diffusion models. Our approach relaxes the original design space into a continuous grid representation, where gradients can be computed by implicit differentiation in the forward simulation. A diffusion model is trained on this relaxed parameter space in order to serve as a prior for plausible relaxed designs. Parameters are sampled by guided diffusion using gradients that are propagated from an objective function specified at inference time through the differentiable simulation. A design sample is obtained by backprojection into the original parameter space. We develop our approach for a composite material design problem where the forward process is modeled as a linear FEM problem. We evaluate the performance of our approach in finding designs that match a specified bulk modulus. We demonstrate that our method can propose diverse designs within 1% relative error margin from medium to high target bulk moduli in 2D and 3D settings. We also demonstrate that the material density of generated samples can be minimized simultaneously by using a multi-objective loss function.

Inverse problems and inverse design in multiphase media, i.e., recovering or engineering microstructures to achieve target macroscopic responses, require operating on discrete-valued material fields, rendering the problem non-differentiable and incompatible with gradient-based methods. Existing approaches either relax to continuous approximations, compromising physical fidelity, or employ separate heavyweight models for forward and inverse tasks. We propose GenPANIS, a unified generative framework that preserves exact discrete microstructures while enabling gradient-based inference through continuous latent embeddings. The model learns a joint distribution over microstructures and PDE solutions, supporting bidirectional inference (forward prediction and inverse recovery) within a single architecture. The generative formulation enables training with unlabeled data, physics residuals, and minimal labeled pairs. A physics-aware decoder incorporating a differentiable coarse-grained PDE solver preserves governing equation structure, enabling extrapolation to varying boundary conditions and microstructural statistics. A learnable normalizing flow prior captures complex posterior structure for inverse problems. Demonstrated on Darcy flow and Helmholtz equations, GenPANIS maintains accuracy on challenging extrapolative scenarios - including unseen boundary conditions, volume fractions, and microstructural morphologies, with sparse, noisy observations. It outperforms state-of-the-art methods while using 10 - 100 times fewer parameters and providing principled uncertainty quantification.

The thermoelastic metamaterial based on a bimaterial hybrid-honeycomb structure, exhibiting simultaneously negative Poisson's ratios and negative thermal expansion coefficients is very promising for various application. This work is dedicated to the machine learning (ML)-enabled inverse design of such structure, on the basis of high-throughput simulation and neural network models. A large dataset is generated through computational homogenization of structures with varying geometrical features and base material properties. A forward ML model is first trained to efficiently and accurately predict the effective thermoelastic properties for a given structure design. Subsequently, inverse ML models are developed to suggest geometrical features and base materials for desired target properties. To address various inverse design scenarios, six different models are proposed, each defined by different combinations of target effective properties and structural design variables. The trained forward model is integrated into the loss functions of the inverse models and is also employed to generate additional datasets for cases with fixed base materials. The good predictive performance of the forward and inverse ML models is demonstrated by representative design examples. These ML models can be applied to efficiently solving specific inverse design tasks involved in the practical application of the thermoelastic metamaterial in novel engineering systems.

When designing new materials, it is often necessary to tailor the material design (with respect to its design parameters) to have some desired properties (e.g. Young's modulus). As the set of design parameters grow, the search space grows exponentially, making the actual synthesis and evaluation of all material combinations virtually impossible. Even using traditional computational methods such as Finite Element Analysis becomes too computationally heavy to search the design space. Recent methods use machine learning (ML) surrogate models to more efficiently determine optimal material designs; unfortunately, these methods often (i) are notoriously difficult to interpret and (ii) under perform when the training data comes from a non-uniform sampling of the design space. We suggest the use of tensor completion methods as an all-in-one approach for interpretability and predictions. We observe classical tensor methods are able to compete with traditional ML in predictions, with the added benefit of their interpretable tensor factors (which are given completely for free, as a result of the prediction). In our experiments, we are able to rediscover physical phenomena via the tensor factors, indicating that our predictions are aligned with the true underlying physics of the problem. This also means these tensor factors could be used by experimentalists to identify potentially novel patterns, given we are able to rediscover existing ones. We also study the effects of both types of surrogate models when we encounter training data from a non-uniform sampling of the design space. We observe more specialized tensor methods that can give better generalization in these non-uniforms sampling scenarios. We find the best generalization comes from a tensor model, which is able to improve upon the baseline ML methods by up to 5% on aggregate $R^2$, and halve the error in some out of distribution regions.

As a pioneer of the third-generation photovoltaic revolution, Perovskite Solar Cells (PSCs) are renowned for their superior optoelectronic performance and cost potential. The development process of PSCs is precise and complex, involving a series of closed-loop workflows such as literature retrieval, data integration, experimental design, and synthesis. However, existing AI perovskite approaches focus predominantly on discrete models, including material design, process optimization,and property prediction. These models fail to propagate physical constraints across the workflow, hindering end-to-end optimization. In this paper, we propose a multi-agent system for perovskite material discovery, named PeroMAS. We first encapsulated a series of perovskite-specific tools into Model Context Protocols (MCPs). By planning and invoking these tools, PeroMAS can design perovskite materials under multi-objective constraints, covering the entire process from literature retrieval and data extraction to property prediction and mechanism analysis. Furthermore, we construct an evaluation benchmark by perovskite human experts to assess this multi-agent system. Results demonstrate that, compared to single Large Language Model (LLM) or traditional search strategies, our system significantly enhances discovery efficiency. It successfully identified candidate materials satisfying multi-objective constraints. Notably, we verify PeroMAS's effectiveness in the physical world through real synthesis experiments.

Electrospinning is a powerful technique for producing micro to nanoscale fibers with application specific architectures. Small variations in solution or operating conditions can shift the jet regime, generating non Gaussian fiber diameter distributions. Despite substantial progress, no existing framework enables inverse design toward desired fiber outcomes while integrating polymer solvent chemical constraints or predicting full distributions. SpinCastML is an open source, distribution aware, chemically informed machine learning and Inverse Monte Carlo (IMC) software for inverse electrospinning design. Built on a rigorously curated dataset of 68,480 fiber diameters from 1,778 datasets across 16 polymers, SpinCastML integrates three structured sampling methods, a suite of 11 high-performance learners, and chemistry aware constraints to predict not only mean diameter but the entire distribution. Cubist model with a polymer balanced Sobol D optimal sampling provides the highest global performance (R2 > 0.92). IMC accurately captures the fiber distributions, achieving R2 > 0.90 and <1% error between predicted and experimental success rates. The IMC engine supports both retrospective analysis and forward-looking inverse design, generating physically and chemically feasible polymer solvent parameter combinations with quantified success probabilities for user-defined targets. SpinCastML reframes electrospinning from trial and error to a reproducible, data driven design process. As an open source executable, it enables laboratories to analyze their own datasets and co create an expanding community software. SpinCastML reduces experimental waste, accelerates discovery, and democratizes access to advanced modeling, establishing distribution aware inverse design as a new standard for sustainable nanofiber manufacturing across biomedical, filtration, and energy applications.

Bayesian optimization (BO) is widely used for autonomous materials discovery, yet its classical sequential formulation is insufficient for design of experimental workflows that often combine parallel or batch synthesis with inherently serial characterization. Methods such as combinatorial spread libraries and printed libraries sample a defined low-D manifold in the chemical space of the system. Here, we introduce a time-aware framework for comparing sequential and manifold BO under experimentally realistic constraints. By explicitly modeling synthesis and characterization times, we define an effective experimental time metric that enables fair, time-normalized benchmarking of optimization strategies. Using numerical experiments in ternary and quaternary compositional spaces, we show that sequential BO remains optimal for short-term experiments or when batching provides no effective time advantage, whereas manifold BO becomes favorable once multiplexed synthesis enables faster accumulation of measurements. We identify a small set of physically interpretable parameters that govern the transition between these regimes. These results establish a general, experimentally grounded framework for selecting optimization strategies in self-driving laboratories and autonomous materials discovery workflows. The accompanying analysis code is publicly available at https://github.com/Slautin/2025_GP_BO_Manifolds.

Large Language Models (LLMs) promise to accelerate discovery by reasoning across the expanding scientific landscape. Yet, the challenge is no longer access to information but connecting it in meaningful, domain-spanning ways. In materials science, where innovation demands integrating concepts from molecular chemistry to mechanical performance, this is especially acute. Neither humans nor single-agent LLMs can fully contend with this torrent of information, with the latter often prone to hallucinations. To address this bottleneck, we introduce a multi-agent framework guided by large-scale knowledge graphs to find sustainable substitutes for per- and polyfluoroalkyl substances (PFAS)-chemicals currently under intense regulatory scrutiny. Agents in the framework specialize in problem decomposition, evidence retrieval, design parameter extraction, and graph traversal, uncovering latent connections across distinct knowledge pockets to support hypothesis generation. Ablation studies show that the full multi-agent pipeline outperforms single-shot prompting, underscoring the value of distributed specialization and relational reasoning. We demonstrate that by tailoring graph traversal strategies, the system alternates between exploitative searches focusing on domain-critical outcomes and exploratory searches surfacing emergent cross-connections. Illustrated through the exemplar of biomedical tubing, the framework generates sustainable PFAS-free alternatives that balance tribological performance, thermal stability, chemical resistance, and biocompatibility. This work establishes a framework combining knowledge graphs with multi-agent reasoning to expand the materials design space, showcasing several initial design candidates to demonstrate the approach.

Generalizing from limited data is particularly critical for models in domains such as material science, where task-relevant features in experimental datasets are often heavily confounded by measurement noise and experimental artifacts. Standard regularization techniques fail to precisely separate meaningful features from noise, while existing adversarial adaptation methods are limited by their reliance on explicit separation labels. To address this challenge, we propose the Adversarial Information Separation Framework (AdverISF), which isolates task-relevant features from noise without requiring explicit supervision. AdverISF introduces a self-supervised adversarial mechanism to enforce statistical independence between task-relevant features and noise representations. It further employs a multi-layer separation architecture that progressively recycles noise information across feature hierarchies to recover features inadvertently discarded as noise, thereby enabling finer-grained feature extraction. Extensive experiments demonstrate that AdverISF outperforms state-of-the-art methods in data-scarce scenarios. In addition, evaluations on real-world material design tasks show that it achieves superior generalization performance.

Generalizing from limited data is particularly critical for models in domains such as material science, where task-relevant features in experimental datasets are often heavily confounded by measurement noise and experimental artifacts. Standard regularization techniques fail to precisely separate meaningful features from noise, while existing adversarial adaptation methods are limited by their reliance on explicit separation labels. To address this challenge, we propose the Adversarial Information Separation Framework (AdverISF), which isolates task-relevant features from noise without requiring explicit supervision. AdverISF introduces a self-supervised adversarial mechanism to enforce statistical independence between task-relevant features and noise representations. It further employs a multi-layer separation architecture that progressively recycles noise information across feature hierarchies to recover features inadvertently discarded as noise, thereby enabling finer-grained feature extraction. Extensive experiments demonstrate that AdverISF outperforms state-of-the-art methods in data-scarce scenarios. In addition, evaluations on real-world material design tasks show that it achieves superior generalization performance.

The design of sustainable materials requires access to materials performance and sustainability data from literature corpus in an organized, structured and automated manner. Natural language processing approaches, particularly large language models (LLMs), have been explored for materials data extraction from the literature, yet often suffer from limited accuracy or narrow scope. In this work, an LLM-based pipeline is developed to accurately extract alloy-related information from both textual descriptions and tabular data across the literature on high-entropy (or multicomponent) alloys (HEA). Specifically two databases with 37,711 and 148,069 entries respectively are retrieved; one from the literature text, consisting of alloy composition, processing conditions, characterization methods, and reported properties, and other from the literature tables, consisting of property names, values, and units. The pipeline enhances materials-domain sensitivity through prompt engineering and retrieval-augmented generation and achieves F1-scores of 0.83 for textual extraction and 0.88 for tabular extraction, surpassing or matching existing approaches. Application of the pipeline to over 10,000 articles yields the largest publicly available multicomponent alloy database and reveals compositional and processing-property trends. The database is further employed for sustainability-aware materials selection in three application domains, i.e., lightweighting, soft magnetic, and corrosion-resistant, identifying multicomponent alloy candidates with more sustainable production while maintaining or exceeding benchmark performance. The pipeline developed can be easily generalized to other class of materials, and assist in development of comprehensive, accurate and usable databases for sustainable materials design.

Inverse design with physics-based objectives is challenging because it couples high-dimensional geometry with expensive simulations, as exemplified by aerodynamic shape optimization for drag reduction. We revisit inverse design through two canonical solutions, the optimal design point and the optimal design distribution, and relate them to optimization and guided generation. Building on this view, we propose a new training loss for cost predictors and a density-gradient optimization method that improves objectives while preserving plausible shapes. We further unify existing training-free guided generation methods. To address their inability to approximate conditional covariance in high dimensions, we develop a time- and memory-efficient algorithm for approximate covariance estimation. Experiments on a controlled 2D study and high-fidelity 3D aerodynamic benchmarks (car and aircraft), validated by OpenFOAM simulations and miniature wind-tunnel tests with 3D-printed prototypes, demonstrate consistent gains in both optimization and guided generation. Additional offline RL results further support the generality of our approach.

Inverse design with physics-based objectives is challenging because it couples high-dimensional geometry with expensive simulations, as exemplified by aerodynamic shape optimization for drag reduction. We revisit inverse design through two canonical solutions, the optimal design point and the optimal design distribution, and relate them to optimization and guided generation. Building on this view, we propose a new training loss for cost predictors and a density-gradient optimization method that improves objectives while preserving plausible shapes. We further unify existing training-free guided generation methods. To address their inability to approximate conditional covariance in high dimensions, we develop a time- and memory-efficient algorithm for approximate covariance estimation. Experiments on a controlled 2D study and high-fidelity 3D aerodynamic benchmarks (car and aircraft), validated by OpenFOAM simulations and miniature wind-tunnel tests with 3D-printed prototypes, demonstrate consistent gains in both optimization and guided generation. Additional offline RL results further support the generality of our approach.

Universal machine-learning interatomic potentials (uMLIPs) have become powerful tools for accelerating computational materials discovery by replacing expensive first-principles calculations in crystal structure prediction (CSP). However, their effectiveness in identifying new, complex materials remains uncertain. Here, we systematically assess the capability of a uMLIP (i.e.,M3GNet) to accelerate CSP in quaternary oxides. Through extensive exploration of the Sr-Li-Al-O and Ba-Y-Al-O systems, we show that uMLIP can rediscover experimentally known materials absent from its training set and identify seven new thermodynamically and dynamically stable compounds. These include a new polymorph of Sr2LiAlO4 (P3221) and a new disordered phase, Sr2Li4Al2O7 (P1_bar). Furthermore, our results show stability predictions based on the semilocal PBE functional require cross-validation with higher-level methods, such as SCAN and RPA, to ensure reliability. While uMLIPs substantially reduce the computational cost of CSP, the primary bottleneck has shifted to the efficiency of search algorithms in navigating complex structural spaces. This work highlights both the promise and current limitations of uMLIP-driven CSP in the discovery of new materials.

Semantic segmentation of microscopy images is a critical task for high-throughput materials characterisation, yet its automation is severely constrained by the prohibitive cost, subjectivity, and scarcity of expert-annotated data. While physics-based simulations offer a scalable alternative to manual labelling, models trained on such data historically fail to generalise due to a significant domain gap, lacking the complex textures, noise patterns, and imaging artefacts inherent to experimental data. This paper introduces a novel framework for labour-free segmentation that successfully bridges this simulation-to-reality gap. Our pipeline leverages phase-field simulations to generate an abundant source of microstructural morphologies with perfect, intrinsically-derived ground-truth masks. We then employ a Cycle-Consistent Generative Adversarial Network (CycleGAN) for unpaired image-to-image translation, transforming the clean simulations into a large-scale dataset of high-fidelity, realistic SEM images. A U-Net model, trained exclusively on this synthetic data, demonstrated remarkable generalisation when deployed on unseen experimental images, achieving a mean Boundary F1-Score of 0.90 and an Intersection over Union (IOU) of 0.88. Comprehensive validation using t-SNE feature-space projection and Shannon entropy analysis confirms that our synthetic images are statistically and featurally indistinguishable from the real data manifold. By completely decoupling model training from manual annotation, our generative framework transforms a data-scarce problem into one of data abundance, providing a robust and fully automated solution to accelerate materials discovery and analysis.

Machine learning (ML) can facilitate efficient thermoelectric (TE) material discovery essential to address the environmental crisis. However, ML models often suffer from poor experimental generalizability despite high metrics. This study presents a robust workflow, applied to the half-Heusler (hH) structural prototype, for figure of merit (zT) prediction, to improve the generalizability of ML models. To resolve challenges in dataset handling and feature filtering, we first introduce a rigorous PCA-based splitting method that ensures training and test sets are unbiased and representative of the full chemical space. We then integrate Bayesian hyperparameter optimization with k-best feature filtering across three architectures-Random Forest, XGBoost, and Neural Networks - while employing SISSO symbolic regression for physical insight and comparison. Using SHAP and SISSO analysis, we identify A-site dopant concentration (xA'), and A-site Heat of Vaporization (HVA) as the primary drivers of zT besides Temperature (T). Finally, a high-throughput screening of approximately 6.6x10^8 potential compositions, filtered by stability constraints, yielded several novel high-zT candidates. Breaking from the traditional focus of improving test RMSE/R^2 values of the models, this work shifts the attention on establishing the test set a true proxy for model generalizability and strengthening the often neglected modules of the existing ML workflows for the data-driven design of next-generation thermoelectric materials.

Designing high-quality, standards-aligned instructional materials for K--12 science is time-consuming and expertise-intensive. This study examines what human experts notice when reviewing AI-generated evaluations of such materials, aiming to translate their insights into design principles for a future GenAI-based instructional material design agent. We intentionally selected 12 high-quality curriculum units across life, physical, and earth sciences from validated programs such as OpenSciEd and Multiple Literacies in Project-based Learning. Using the EQuIP rubric with 9 evaluation items, we prompted GPT-4o, Claude, and Gemini to produce numerical ratings and written rationales for each unit, generating 648 evaluation outputs. Two science education experts independently reviewed all outputs, marking agreement (1) or disagreement (0) for both scores and rationales, and offering qualitative reflections on AI reasoning. This process surfaces patterns in where LLM judgments align with or diverge from expert perspectives, revealing reasoning strengths, gaps, and contextual nuances. These insights will directly inform the development of a domain-specific GenAI agent to support the design of high-quality instructional materials in K--12 science education.

Continuous-time generative models for crystalline materials enable inverse materials design by learning to predict stable crystal structures, but incorporating explicit target properties into the generative process remains challenging. Policy-gradient reinforcement learning (RL) provides a principled mechanism for aligning generative models with downstream objectives but typically requires access to the score, which has prevented its application to flow-based models that learn only velocity fields. We introduce Open Materials Generation with Inference-time Reinforcement Learning (OMatG-IRL), a policy-gradient RL framework that operates directly on the learned velocity fields and eliminates the need for the explicit computation of the score. OMatG-IRL leverages stochastic perturbations of the underlying generation dynamics preserving the baseline performance of the pretrained generative model while enabling exploration and policy-gradient estimation at inference time. Using OMatG-IRL, we present the first application of RL to crystal structure prediction (CSP). Our method enables effective reinforcement of an energy-based objective while preserving diversity through composition conditioning, and it achieves performance competitive with score-based RL approaches. Finally, we show that OMatG-IRL can learn time-dependent velocity-annealing schedules, enabling accurate CSP with order-of-magnitude improvements in sampling efficiency and, correspondingly, reduction in generation time.

Generative learning generates high dimensional data based on low dimensional conditions, also called prompts. Therefore, generative learning algorithms are eligible for solving (Bayesian) inverse problems. In this article we compare a traditional Bayesian inverse approach based on a forward regression model and a prior sampled with the Markov Chain Monte Carlo method with three state of the art generative learning models, namely conditional Generative Adversarial Networks, Invertible Neural Networks and Conditional Flow Matching. We apply them to a problem of gas turbine combustor design where we map six independent design parameters to three performance labels. We propose several metrics for the evaluation of this inverse design approaches and measure the accuracy of the labels of the generated designs along with the diversity. We also study the performance as a function of the training dataset size. Our benchmark has a clear winner, as Conditional Flow Matching consistently outperforms all competing approaches.

In this work, we present Multimodal Equivariant Inverse Design Network (MEIDNet), a framework that jointly learns structural information and materials properties through contrastive learning, while encoding structures via an equivariant graph neural network (EGNN). By combining generative inverse design with multimodal learning, our approach accelerates the exploration of chemical-structural space and facilitates the discovery of materials that satisfy predefined property targets. MEIDNet exhibits strong latent-space alignment with cosine similarity 0.96 by fusion of three modalities through cross-modal learning. Through implementation of curriculum learning strategies, MEIDNet achieves ~60 times higher learning efficiency than conventional training techniques. The potential of our multimodal approach is demonstrated by generating low-bandgap perovskite structures at a stable, unique, and novel (SUN) rate of 13.6 %, which are further validated by ab initio methods. Our inverse design framework demonstrates both scalability and adaptability, paving the way for the universal learning of chemical space across diverse modalities.

Artificial intelligence (AI) has transformed materials discovery, enabling rapid exploration of chemical space through generative models and surrogate screening. Yet current AI workflows optimize performance first, deferring sustainability to post synthesis assessment. This creates inefficiency by the time environmental burdens are quantified, resources have been invested in potentially unsustainable solutions. The disconnect between atomic scale design and lifecycle assessment (LCA) reflects fundamental challenges, data scarcity across heterogeneous sources, scale gaps from atoms to industrial systems, uncertainty in synthesis pathways, and the absence of frameworks that co-optimize performance with environmental impact. We propose to integrate upstream machine learning (ML) assisted materials discovery with downstream lifecycle assessment into a uniform ML-LCA environment. The framework ML-LCA integrates five components, information extraction for building materials-environment knowledge bases, harmonized databases linking properties to sustainability metrics, multi-scale models bridging atomic properties to lifecycle impacts, ensemble prediction of manufacturing pathways with uncertainty quantification, and uncertainty-aware optimization enabling simultaneous performance-sustainability navigation. Case studies spanning glass, cement, semiconductor photoresists, and polymers demonstrate both necessity and feasibility while identifying material-specific integration challenges. Realizing ML-LCA demands coordinated advances in data infrastructure, ex-ante assessment methodologies, multi-objective optimization, and regulatory alignment enabling the discovery of materials that are sustainable by design rather than by chance.

Artificial intelligence (AI) has transformed materials discovery, enabling rapid exploration of chemical space through generative models and surrogate screening. Yet current AI workflows optimize performance first, deferring sustainability to post synthesis assessment. This creates inefficiency by the time environmental burdens are quantified, resources have been invested in potentially unsustainable solutions. The disconnect between atomic scale design and lifecycle assessment (LCA) reflects fundamental challenges, data scarcity across heterogeneous sources, scale gaps from atoms to industrial systems, uncertainty in synthesis pathways, and the absence of frameworks that co-optimize performance with environmental impact. We propose to integrate upstream machine learning (ML) assisted materials discovery with downstream lifecycle assessment into a uniform ML-LCA environment. The framework ML-LCA integrates five components, information extraction for building materials-environment knowledge bases, harmonized databases linking properties to sustainability metrics, multi-scale models bridging atomic properties to lifecycle impacts, ensemble prediction of manufacturing pathways with uncertainty quantification, and uncertainty-aware optimization enabling simultaneous performance-sustainability navigation. Case studies spanning glass, cement, semiconductor photoresists, and polymers demonstrate both necessity and feasibility while identifying material-specific integration challenges. Realizing ML-LCA demands coordinated advances in data infrastructure, ex-ante assessment methodologies, multi-objective optimization, and regulatory alignment enabling the discovery of materials that are sustainable by design rather than by chance.

Electrides are exotic compounds in which excess electrons occupy interstitial regions of the crystal lattice and serve as anions, exhibiting exceptional properties such as low work function, high electron mobility, and strong catalytic activity. Although they show promise for diverse applications, identifying new electrides remains challenging due to the difficulty of achieving energetically favorable electron localization in crystal cavities. Here, we present an accelerated materials discovery framework that combines physical principles, diffusion-based materials generation with hierarchical thermodynamic and electronic structure screening. Using this workflow, we systematically explored 1,510 binary and 6,654 ternary chemical compositions containing excess valence electrons from electropositive alkaline, alkaline-earth, and early transition metals, and then filtered them with a high throughput validation on both thermodynamical stability and electronic structure analysis. As a result, we have identified 264 new electron rich compounds within 0.05 eV/atom above the convex hull at the density functional theory (DFT) level, including 13 thermodynamically stable electrides. Our approach demonstrates a generalizable strategy for targeted materials discovery in a vast chemical space.

Superconducting materials are of significant technological relevance for a broad range of applications, and intense research efforts aim at enhancing the critical temperature $T_{c}$. Intriguingly, while numerous studies have explored different computational and machine-learning routes to predict $T_{c}$, the fundamental role of the critical magnetic field has so far been overlooked. Here we open a new frontier in superconductor discovery by presenting a consistent computational database of critical fields $H_{c}$, $H_{c1}$, and $H_{c2}$ for over 7300 electron-phonon-paired superconductors covering distinct materials classes. A theoretical framework is developed that combines $α^2F(ω)$ spectral functions and highly accurate Fermi surfaces from density functional theory with clean-limit Eliashberg theory to obtain the coherence lengths, London penetration depths, and Ginzburg-Landau parameters. We discover an unexpectedly large number of Type-I superconductors and show that larger unit cells generically support higher critical fields and Type-II behavior. We identify the importance of going beyond BCS theory by including strong-coupling corrections to the superconducting gap and electron-phonon renormalizations of the effective mass for predictions of critical fields across materials. These results provide a framework for foundational AI models that realize the concept of inverse materials design for high-$T_{c}$ and high-critical-field superconductors.

Existing benchmarks for computational materials discovery primarily evaluate static predictive tasks or isolated computational sub-tasks. While valuable, these evaluations neglect the inherently iterative and adaptive nature of scientific discovery. We introduce MAterials Discovery Environments (MADE), a novel framework for benchmarking end-to-end autonomous materials discovery pipelines. MADE simulates closed-loop discovery campaigns in which an agent or algorithm proposes, evaluates, and refines candidate materials under a constrained oracle budget, capturing the sequential and resource-limited nature of real discovery workflows. We formalize discovery as a search for thermodynamically stable compounds relative to a given convex hull, and evaluate efficacy and efficiency via comparison to baseline algorithms. The framework is flexible; users can compose discovery agents from interchangeable components such as generative models, filters, and planners, enabling the study of arbitrary workflows ranging from fixed pipelines to fully agentic systems with tool use and adaptive decision making. We demonstrate this by conducting systematic experiments across a family of systems, enabling ablation of components in discovery pipelines, and comparison of how methods scale with system complexity.

Developing new metal hydrides is a critical step toward efficient hydrogen storage in carbon-neutral energy systems. However, existing materials databases, such as the Materials Project, contain a limited number of well-characterized hydrides, which constrains the discovery of optimal candidates. This work presents a framework that integrates causal discovery with a lightweight generative machine learning model to generate novel metal hydride candidates that may not exist in current databases. Using a dataset of 450 samples (270 training, 90 validation, and 90 testing), the model generates 1,000 candidates. After ranking and filtering, six previously unreported chemical formulas and crystal structures are identified, four of which are validated by density functional theory simulations and show strong potential for future experimental investigation. Overall, the proposed framework provides a scalable and time-efficient approach for expanding hydrogen storage datasets and accelerating materials discovery.

Generative models for materials have achieved strong performance on periodic bulk crystals, yet their ability to generalize across scale transitions to finite nanostructures remains largely untested. We introduce Crystal-to-Nanoparticle (C2NP), a systematic benchmark for evaluating generative models when moving between infinite crystalline unit cells and finite nanoparticles, where surface effects and size-dependent distortions dominate. C2NP defines two complementary tasks: (i) generating nanoparticles of specified radii from periodic unit cells, testing whether models capture surface truncation and geometric constraints; and (ii) recovering bulk lattice parameters and space-group symmetry from finite particle configurations, assessing whether models can infer underlying crystallographic order despite surface perturbations. Using diverse materials as a structurally consistent testbed, we construct over 170,000 nanoparticle configurations by carving particles from supercells derived from DFT-relaxed crystal unit cells, and introduce size-based splits that separate interpolation from extrapolation regimes. Experiments with state-of-the-art approaches, including diffusion, flow-matching, and variational models, show that even when losses are low, models often fail geometrically under distribution shift, yielding large lattice-recovery errors and near-zero joint accuracy on structure and symmetry. Overall, our results suggest that current methods rely on template memorization rather than scalable physical generalization. C2NP offers a controlled, reproducible framework for diagnosing these failures, with immediate applications to nanoparticle catalyst design, nanostructured hydrides for hydrogen storage, and materials discovery. Dataset and code are available at https://github.com/KurbanIntelligenceLab/C2NP.

Water electrolysis is an eco-friendly method for hydrogen production that has reached significant levels of technological maturity. Among commercialized water-electrolysis technologies, proton-exchange membrane electrolyzers offer high current density, fast dynamic response, and compact system design, among other advantages. On the other hand, managing their high capital cost and the ``forever-chemistry'' nature of Nafion, a perfluorinated proton-exchange membrane widely used in such devices, remains a major challenge. Searches for fluorine-free replacements for Nafion, pursued largely through physical experimentation, have been active for decades with limited success. In this work, we develop and demonstrate an AI-based strategy for designing new proton-exchange membranes for electrolyzers. Two key components of this strategy are an implementation of the virtual forward-synthesis approach and a set of machine-learning predictive models for essential application-inspired membrane properties; the former generates a vast space of millions of synthesizable polymers, which are then evaluated and screened by the latter. The strategy is validated against experimental data for known membranes and then applied to design over 1,700 new synthesizable candidates. This article concludes with a forward-looking vision in which the strategy could be elevated into an interactive and iterative scheme that are based on large language models to facilitate materials design in multiple ways.

Artificial intelligence (AI) has shown remarkable success in materials discovery and property prediction, particularly for crystalline and polymer systems where material properties and structures are dominated by discrete graph representations. Such graph-central paradigm breaks down on composite materials, which possess continuous and nonlinear design spaces that lack well-defined graph structures. General composite descriptors, e.g., fiber volume and misalignment angle, cannot fully capture the fiber distributions that fundamentally determine microstructural characteristics, necessitating the integration of heterogeneous data sources through multimodal learning. Existing alignment-oriented multimodal frameworks have proven effective on abundant crystal or polymer data under discrete, unique graph-property mapping assumptions, but fail to address the highly continuous composite design space under extreme data scarcity. In this work, we introduce ORDinal-aware imagE-tabulaR alignment (ORDER), a multimodal pretraining framework that establishes ordinality as a core principle for composite material representations. ORDER ensures that materials with similar target properties occupy nearby regions in the latent space, which effectively preserves the continuous nature of composite properties and enables meaningful interpolation between sparsely observed designs. We evaluate ORDER on a public Nanofiber-enforced composite dataset and an internally curated dataset that simulates the construction of carbon fiber T700 with diverse fiber distributions. ORDER achieves consistent improvements over state-of-the-art multimodal baselines across property prediction, cross-modal retrieval, and microstructure generation tasks.

Active learning for photonic crystals explores the integration of analytic approximate Bayesian last layer neural networks (LL-BNNs) with uncertainty-driven sample selection to accelerate photonic band gap prediction. We employ an analytic LL-BNN formulation, corresponding to the infinite Monte Carlo sample limit, to obtain uncertainty estimates that are strongly correlated with the true predictive error on unlabeled candidate structures. These uncertainty scores drive an active learning strategy that prioritizes the most informative simulations during training. Applied to the task of predicting band gap sizes in two-dimensional, two-tone photonic crystals, our approach achieves up to a 2.6x reduction in required training data compared to a random sampling baseline while maintaining predictive accuracy. The efficiency gains arise from concentrating computational resources on high uncertainty regions of the design space rather than sampling uniformly. Given the substantial cost of full band structure simulations, especially in three dimensions, this data efficiency enables rapid and scalable surrogate modeling. Our results suggest that analytic LL-BNN based active learning can substantially accelerate topological optimization and inverse design workflows for photonic crystals, and more broadly, offers a general framework for data efficient regression across scientific machine learning domains.

We propose Materealize, a multi-agent system for end-to-end inorganic materials design and synthesis that orchestrates core domain tools spanning structure generation, property prediction, synthesizability prediction, and synthesis planning within a single unified framework. Through a natural-language interface, Materealize enables non-experts to access computational materials workflows and obtain experimentally actionable outputs for material realization. Materealize provides two complementary modes. In instant mode, the system rapidly composes connected tools to solve diverse inorganic tasks-including property-conditioned synthesizable candidate design with synthesis recipes, diagnosis, and redesign of unsynthesizable structures, and synthesizable data augmentation-within a few minutes. In thinking mode, Materealize applies multi-agent debate to deliver more refined and information-rich synthesis recommendations, including reasoning- and model-driven synthesis routes and mechanistic hypotheses. The mechanistic hypotheses are validated by direct comparison with the literature for known mechanisms and further supported by physics-grounded simulations for novel synthesis pathways. By combining tool-level accuracy with reasoning-level integration, Materealize can bridge the gap between computational discovery and practical experimental realization.

We introduce a physics-informed Generative Adversarial Network framework that recasts quantum resource-state generation as an inverse-design task. By embedding task-specific utility functions into training, the model learns to generate valid two-qubit states optimized for teleportation and entanglement broadcasting. Comparing decomposition-based and direct-generation architectures reveals that structural enforcement of Hermiticity, trace-one, and positivity yields higher fidelity and training stability than loss-only approaches. The framework reproduces theoretical resource boundaries for Werner-like and Bell-diagonal states with fidelities exceeding ~98%, establishing adversarial learning as a lightweight yet effective method for constraint-driven quantum-state discovery. This approach provides a scalable foundation for automated design of tailored quantum resources for information-processing applications, exemplified with teleportation and broadcasting of entanglement, and it opens up the possibility of using such states in efficient quantum network design.

We introduce a physics-informed Generative Adversarial Network framework that recasts quantum resource-state generation as an inverse-design task. By embedding task-specific utility functions into training, the model learns to generate valid two-qubit states optimized for teleportation and entanglement broadcasting. Comparing decomposition-based and direct-generation architectures reveals that structural enforcement of Hermiticity, trace-one, and positivity yields higher fidelity and training stability than loss-only approaches. The framework reproduces theoretical resource boundaries for Werner-like and Bell-diagonal states with fidelities exceeding ~98%, establishing adversarial learning as a lightweight yet effective method for constraint-driven quantum-state discovery. This approach provides a scalable foundation for automated design of tailored quantum resources for information-processing applications, exemplified with teleportation and broadcasting of entanglement, and it opens up the possibility of using such states in efficient quantum network design.

Designing fluorescent small molecules with tailored optical and physicochemical properties requires navigating vast, underexplored chemical space while satisfying multiple objectives and constraints. Conventional generate-score-screen approaches become impractical under such realistic design specifications, owing to their low search efficiency, unreliable generalizability of machine-learning prediction, and the prohibitive cost of quantum chemical calculation. Here we present LUMOS, a data-and-physics driven framework for inverse design of fluorescent molecules. LUMOS couples generator and predictor within a shared latent representation, enabling direct specification-to-molecule design and efficient exploration. Moreover, LUMOS combines neural networks with a fast time-dependent density functional theory (TD-DFT) calculation workflow to build a suite of complementary predictors spanning different trade-offs in speed, accuracy, and generalizability, enabling reliable property prediction across diverse scenarios. Finally, LUMOS employs a property-guided diffusion model integrated with multi-objective evolutionary algorithms, enabling de novo design and molecular optimization under multiple objectives and constraints. Across comprehensive benchmarks, LUMOS consistently outperforms baseline models in terms of accuracy, generalizability and physical plausibility for fluorescence property prediction, and demonstrates superior performance in multi-objective scaffold- and fragment-level molecular optimization. Further validation using TD-DFT and molecular dynamics (MD) simulations demonstrates that LUMOS can generate valid fluorophores that meet various target specifications. Overall, these results establish LUMOS as a data-physics dual-driven framework for general fluorophore inverse design.

Metamaterials design for advanced functionality often entails the inverse design on nonlinear and condition-dependent responses (e.g., stress-strain relation and dispersion relation), which are described by continuous functions. Most existing design methods focus on vector-valued responses (e.g., Young's modulus and bandgap width), while the inverse design of functional responses remains challenging due to their high-dimensionality, the complexity of accommodating design requirements in inverse-design frameworks, and non-existence or non-uniqueness of feasible solutions. Although generative design approaches have shown promise, they are often data-hungry, handle design requirements heuristically, and may generate infeasible designs without uncertainty quantification. To address these challenges, we introduce a RAndom-forest-based Generative approach (RAG). By leveraging the small-data compatibility of random forests, RAG enables data-efficient predictions of high-dimensional functional responses. During the inverse design, the framework estimates the likelihood through the ensemble which quantifies the trustworthiness of generated designs while reflecting the relative difficulty across different requirements. The one-to-many mapping is addressed through single-shot design generation by sampling from the conditional likelihood. We demonstrate RAG on: 1) acoustic metamaterials with prescribed partial passbands/stopbands, and 2) mechanical metamaterials with targeted snap-through responses, using 500 and 1057 samples, respectively. Its data-efficiency is benchmarked against neural networks on a public mechanical metamaterial dataset with nonlinear stress-strain relations. Our framework provides a lightweight, trustworthy pathway to inverse design involving functional responses, expensive simulations, and complex design requirements, beyond metamaterials.

Artificial Intelligence is rapidly transforming materials science and engineering, offering powerful tools to navigate complexity, accelerate discovery, and optimize material design in ways previously unattainable. Driven by the accelerating pace of algorithmic advancements and increasing data availability, AI is becoming an essential competency for materials researchers. This review provides a comprehensive and structured overview of the current landscape, synthesizing recent advancements and methodologies for materials scientists seeking to effectively leverage these data-driven techniques. We survey the spectrum of machine learning approaches, from traditional algorithms to advanced deep learning architectures, including CNNs, GNNs, and Transformers, alongside emerging generative AI and probabilistic models such as Gaussian Processes for uncertainty quantification. The review also examines the pivotal role of data in this field, emphasizing how effective representation and featurization strategies, spanning compositional, structural, image-based, and language-inspired approaches, combined with appropriate preprocessing, fundamentally underpin the performance of machine learning models in materials research. Persistent challenges related to data quality, quantity, and standardization, which critically impact model development and application in materials science and engineering, are also addressed.

Inverse design tools such as Topology Optimization (TO) can achieve new levels of improvement for high-performance engineered structures. However, widespread use is hindered by high computational times and a black-box nature that inhibits user interaction. Human-in-the-loop TO approaches are emerging that integrate human intuition into the design generation process. However, these rely on the time-consuming bottleneck of iterative region selection for design modifications. To reduce the number of iterative trials, this contribution presents an AI co-pilot that uses machine learning to predict the user's preferred regions. The prediction model is configured as an image segmentation task with a U-Net architecture. It is trained on synthetic datasets where human preferences either identify the longest topological member or the most complex structural connection. The model successfully predicts plausible regions for modification and presents them to the user as AI recommendations. The human preference model demonstrates generalization across diverse and non-standard TO problems and exhibits emergent behavior outside the single-region selection training data. Demonstration examples show that the new human-in-the-loop TO approach that integrates the AI co-pilot can improve manufacturability or improve the linear buckling load by 39% while only increasing the total design time by 15 sec compared to conventional simplistic TO.

Hihg-temperature creep characterization of structural alloys traditionally relies on serial uniaxial tests, which are highly inefficient for exploring the large search space of alloy compositions and for material discovery. Here, we introduce a machine-learning-assisted, high-throughput framework for creep law identification based on a dimple array bulge instrument (DABI) configuration, which enables parallel creep testing of 25 dimples, each fabricated from a different alloy, in a single experiment. Full-field surface displacements of dimples undergoing time-dependent creep-induced bulging under inert gas pressure are measured by 3D digital image correlation. We train a recurrent neural network (RNN) as a surrogate model, mapping creep parameters and loading conditions to the time-dependent deformation response of DABI. Coupling this surrogate with a particle swarm optimization scheme enables rapid and global inverse identification with sparsity regularization of creep parameters from experiment displacement-time histories. In addition, we propose a phenomenological creep law with a time-dependent stress exponent that captures the sigmoidal primary creep observed in wrought INCONEL 625 and extracts its temperature dependence from DABI test at multiple temperatures. Furthermore, we employ a general creep law combining several conventional forms together with regularized inversion to identify the creep laws for 47 additional Fe-, Ni-, and Co-rich alloys and to automatically select the dominant functional form for each alloy. This workflow combined with DABI experiment provides a quantitative, high-throughput creep characterization platform that is compatible with data mining, composition-property modeling, and nonlinear structural optimization with creep behavior across a large alloy design space.

The application of generative models in crystal structure prediction (CSP) has gained significant attention. Conditional generation--particularly the generation of crystal structures with specified stability or other physical properties has been actively researched for material discovery purposes. Meanwhile, the generative models capable of symmetry-aware generation are also under active development, because space group symmetry has a strong relationship with the physical properties of materials. In this study, we demonstrate that the symmetry control in the previous conditional crystal generation model may not be sufficiently effective when space group constraints are applied as a condition. To address this problem, we propose the WyckoffDiff-Adaptor, which embeds conditional generation within a WyckoffDiff architecture that effectively diffuses Wyckoff positions to achieve precise symmetry control. We successfully generated formation energy phase diagrams while specifying stable structures of particular combination of elements, such as Li--O and Ti--O systems, while simultaneously preserving the symmetry of the input conditions. The proposed method with symmetry-aware conditional generation demonstrates promising results as an effective approach to achieving the discovery of novel materials with targeted physical properties.

The acceleration of materials discovery requires digital platforms that go beyond data repositories to embed learning, optimization, and decision-making directly into research workflows. We introduce DataScribe, an AI-native, cloud-based materials discovery platform that unifies heterogeneous experimental and computational data through ontology-backed ingestion and machine-actionable knowledge graphs. The platform integrates FAIR-compliant metadata capture, schema and unit harmonization, uncertainty-aware surrogate modeling, and native multi-objective multi-fidelity Bayesian optimization, enabling closed-loop propose-measure-learn workflows across experimental and computational pipelines. DataScribe functions as an application-layer intelligence stack, coupling data governance, optimization, and explainability rather than treating them as downstream add-ons. We validate the platform through case studies in electrochemical materials and high-entropy alloys, demonstrating end-to-end data fusion, real-time optimization, and reproducible exploration of multi-objective trade spaces. By embedding optimization engines, machine learning, and unified access to public and private scientific data directly within the data infrastructure, and by supporting open, free use for academic and non-profit researchers, DataScribe functions as a general-purpose application-layer backbone for laboratories of any scale, including self-driving laboratories and geographically distributed materials acceleration platforms, with built-in support for performance, sustainability, and supply-chain-aware objectives.

Materials discovery is a cornerstone of modern technological advancement, yet it remains constrained by traditional trial-and-error paradigms and the inherent bias of human intuition. Artificial intelligence (AI) has emerged as a transformative tool in materials science by effectively modeling structure-property relationships. Despite substantial efforts to enhance model expressiveness, data efficiency remains an equally critical challenge, given the limited availability of experimental and computational resources. Active learning (AL), as a data-driven machine learning paradigm, has shown great promise for discovering novel materials and enabling the efficient navigation of vast materials spaces. In this review, we follow the evolution of sampling strategy design techniques in AL, from Bayesian optimization to advanced deep learning-based strategies. We then highlight how AL enhances data efficiency across various data regimes, ranging from task-specific settings with limited data to the development of general-purpose datasets and large-scale models. We further provide a systematic overview of AL applications throughout the materials research pipeline, including computational simulation, composition and structural design, process optimization, and self-driving laboratory systems. Finally, we pinpoint key challenges and future perspectives of AL in materials discovery.

A central problem in machine learning theory is to characterize how learning dynamics select particular solutions among the many compatible with the training objective, a phenomenon, called implicit bias, which remains only partially characterized. In the present work, we identify a general mechanism, in terms of an explicit geometric correction of the learning dynamics, for the emergence of implicit biases, arising from the interaction between continuous symmetries in the model's parametrization and stochasticity in the optimization process. Our viewpoint is constructive in two complementary directions: given model symmetries, one can derive the implicit bias they induce; conversely, one can inverse-design a wide class of different implicit biases by computing specific redundant parameterizations. More precisely, we show that, when the dynamics is expressed in the quotient space obtained by factoring out the symmetry group of the parameterization, the resulting stochastic differential equation gains a closed form geometric correction in the stationary distribution of the optimizer dynamics favoring orbits with small local volume. We compute the resulting symmetry induced bias for a range of architectures, showing how several well known results fit into a single unified framework. The approach also provides a practical methodology for deriving implicit biases in new settings, and it yields concrete, testable predictions that we confirm by numerical simulations on toy models trained on synthetic data, leaving more complex scenarios for future work. Finally, we test the implicit bias inverse-design procedure in notable cases, including biases toward sparsity in linear features or in spectral properties of the model parameters.

Combinatorial materials libraries are an efficient route to generate large families of candidate compositions, but their impact is often limited by the speed and depth of characterization and by the difficulty of extracting actionable structure-property relations from complex characterization data. Here we develop an autonomous scanning probe microscopy (SPM) framework that integrates automated atomic force and magnetic force microscopy (AFM/MFM) to rapidly explore magnetic and structural properties across combinatorial spread libraries. To enable automated exploration of systems without a clear optimization target, we introduce a combination of a static physics-informed bag-of-features (BoF) representation of measured surface morphology and magnetic structure with multi-objective Bayesian optimization (MOBO) to discover the relative significance and robustness of features. The resulting closed-loop workflow selectively samples the compositional gradient and reconstructs feature landscapes consistent with dense grid "ground truth" measurements. The resulting Pareto structure reveals where multiple nanoscale objectives are simultaneously optimized, where trade-offs between roughness, coherence, and magnetic contrast are unavoidable, and how families of compositions cluster into distinct functional regimes, thereby turning multi-feature imaging data into interpretable maps of competing structure-property trends. While demonstrated for Au-Co-Ni and AFM/MFM, the approach is general and can be extended to other combinatorial systems, imaging modalities, and feature sets, illustrating how feature-based MOBO and autonomous SPM can transform microscopy images from static data products into active feedback for real-time, multi-objective materials discovery.

Designing mechanical linkages to achieve target end-effector trajectories presents a fundamental challenge due to the intricate coupling between continuous node placements, discrete topological configurations, and nonlinear kinematic constraints. The highly nonlinear motion-to-configuration relationship means small perturbations in joint positions drastically alter trajectories, while the combinatorially expanding design space renders conventional optimization and heuristic methods computationally intractable. We introduce an autoregressive diffusion framework that exploits the dyadic nature of linkage assembly by representing mechanisms as sequentially constructed graphs, where nodes correspond to joints and edges to rigid links. Our approach combines a causal transformer with a Denoising Diffusion Probabilistic Model (DDPM), both conditioned on target trajectories encoded via a transformer encoder. The causal transformer autoregressively predicts discrete topology node-by-node, while the DDPM refines each node's spatial coordinates and edge connectivity to previously generated nodes. This sequential generation enables adaptive trial-and-error synthesis where problematic nodes exhibiting kinematic locking or collisions can be selectively regenerated, allowing autonomous correction of degenerate configurations during design. Our graph-based, data-driven methodology surpasses traditional optimization approaches, enabling scalable inverse design that generalizes to mechanisms with arbitrary node counts. We demonstrate successful synthesis of linkage systems containing up to 20 nodes with extensibility to N-node architectures. This work advances autoregressive graph generation methodologies and computational kinematic synthesis, establishing new paradigms for scalable inverse design of complex mechanical systems.

In computational physics and materials science, first-principles methods, particularly density functional theory, have become central tools for electronic structure prediction and materials design. Recently, rapid advances in artificial intelligence (AI) have begun to reshape the research landscape, giving rise to the emerging field of deep-learning electronic structure calculations. Despite numerous pioneering studies, the field remains in its early stages; existing software implementations are often fragmented, lacking unified frameworks and standardized interfaces required for broad community adoption. Here we present DeepH-pack, a comprehensive and unified software package that integrates first-principles calculations with deep learning. By incorporating fundamental physical principles into neural-network design, such as the nearsightedness principle and the equivariance principle, DeepH-pack achieves robust cross-scale and cross-material generalizability. This allows models trained on small-scale structures to generalize to large-scale and previously unseen materials. The toolkit preserves first-principles accuracy while accelerating electronic structure calculations by several orders of magnitude, establishing an efficient and intelligent computational paradigm for large-scale materials simulation, high-throughput materials database construction, and AI-driven materials discovery.

The synthesis of nanocrystals has been highly dependent on trial-and-error, due to the complex correlation between synthesis parameters and physicochemical properties. Although deep learning offers a potential methodology to achieve generative inverse design, it is still hindered by the scarcity of high-quality datasets that align nanocrystal synthesis routes with their properties. Here, we present the construction of a large-scale, aligned Nanocrystal Synthesis-Property (NSP) database and demonstrate its capability for generative inverse design. To extract structured synthesis routes and their corresponding product properties from literature, we develop NanoExtractor, a large language model (LLM) enhanced by well-designed augmentation strategies. NanoExtractor is validated against human experts, achieving a weighted average score of 88% on the test set, significantly outperforming chemistry-specialized (3%) and general-purpose LLMs (38%). The resulting NSP database contains nearly 160,000 aligned entries and serves as training data for our NanoDesigner, an LLM for inverse synthesis design. The generative capability of NanoDesigner is validated through the successful design of viable synthesis routes for both well-established PbSe nanocrystals and rarely reported MgF2 nanocrystals. Notably, the model recommends a counter-intuitive, non-stoichiometric precursor ratio (1:1) for MgF2 nanocrystals, which is experimentally confirmed as critical for suppressing byproducts. Our work bridges the gap between unstructured literature and data-driven synthesis, and also establishes a powerful human-AI collaborative paradigm for accelerating nanocrystal discovery.

Photonics is becoming a cornerstone technology for high-performance AI systems and scientific computing, offering unparalleled speed, parallelism, and energy efficiency. Despite this promise, the design and deployment of electronic-photonic AI systems remain highly challenging due to a steep learning curve across multiple layers, spanning device physics, circuit design, system architecture, and AI algorithms. The absence of a mature electronic-photonic design automation (EPDA) toolchain leads to long, inefficient design cycles and limits cross-disciplinary innovation and co-evolution. In this work, we present a cross-layer co-design and automation framework aimed at democratizing photonic AI system development. We begin by introducing our architecture designs for scalable photonic edge AI and Transformer inference, followed by SimPhony, an open-source modeling tool for rapid EPIC AI system evaluation and design-space exploration. We then highlight advances in AI-enabled photonic design automation, including physical AI-based Maxwell solvers, a fabrication-aware inverse design framework, and a scalable inverse training algorithm for meta-optical neural networks, enabling a scalable EPDA stack for next-generation electronic-photonic AI systems.

Generative models hold great promise for accelerating materials discovery, but their evaluation often overlooks the chemical validity and stability requirements crucial to real-world applications. Density Functional Theory (DFT) simulations are the gold standard for evaluating such properties but are computationally intensive and inaccessible to non-experts. We propose a chemically grounded, user-friendly evaluation framework that integrates DFT-based stability analysis with commonly used machine learning (ML) metrics. Through systematic experiments using both perturbative and generative methods, we demonstrate that conventional ML metrics can misrepresent chemical feasibility. To address this, we propose new insights on robust metrics and highlight the importance of simulation-informed evaluation for developing reliable generative models in materials science.

Adsorption energy is a key descriptor of catalytic reactivity. It is fundamentally defined as the difference between the relaxed total energy of the adsorbate-surface system and that of an appropriate reference state; therefore, the accuracy of relaxed-energy prediction directly determines the reliability of machine-learning-driven catalyst screening. E(3)-equivariant graph neural networks (GNNs) can natively operate on three-dimensional atomic coordinates under periodic boundary conditions and have demonstrated strong performance on such tasks. In contrast, language-model-based approaches, while enabling human-readable textual descriptions and reducing reliance on explicit graph -- thereby broadening applicability -- remain insufficient in both adsorption-configuration energy prediction accuracy and in distinguishing ``the same system with different configurations,'' even with graph-assisted pretraining in the style of GAP-CATBERTa. To this end, we propose QE-Catalytic, a multimodal framework that deeply couples a large language model (\textbf{Q}wen) with an E(3)-equivariant graph Transformer (\textbf{E}quiformer-V2), enabling unified support for adsorption-configuration property prediction and inverse design on complex catalytic surfaces. During prediction, QE-Catalytic jointly leverages three-dimensional structures and structured configuration text, and injects ``3D geometric information'' into the language channel via graph-text alignment, allowing it to function as a high-performance text-based predictor when precise coordinates are unavailable, while also autoregressively generating CIF files for target-energy-driven structure design and information completion. On OC20, QE-Catalytic reduces the MAE of relaxed adsorption energy from 0.713~eV to 0.486~eV, and consistently outperforms baseline models such as CatBERTa and GAP-CATBERTa across multiple evaluation protocols.

This study aims to evaluate the optoelectronic properties of metal free porphyrin-based D-$π$-A dyes via in-silico performance investigation notifying energy informatics and decision support. To develop novel organic dyes, three acceptor/anchoring groups and five donating groups were introduced to strategic positions of the base porphyrin structure, resulting in a total of fifteen dyes. The singlet ground state geometries of the dyes were optimized utilizing density functional theory (DFT) with B3LYP and the excited state optical properties were explored through time-dependent DFT (TD-DFT) using the PCM model with tetrahydrofuran (THF) as solvent. Both DFT and TD-DFT calculations were carried out using the 6-311G(d,p) basis set. The HOMO energy levels of almost all the modified dyes are lower than the redox potential of I$^-$/I$3^-$ and LUMO energy levels are higher than the conduction band of TiO$2$. The absorption maxima values ranged from 690.64 to 975.55 nm. The dye N1 using triphenylamine group as donor and p-ethynylbenzoic acid group as acceptor, showed optimum optoelectronic properties ($ΔG{reg}=-9.73$ eV, $ΔG{inj}=7.18$ eV, $V_{OC}=1.47$ V and $J_{SC}=15.03$ mA/cm$^2$) with highest PCE 14.37%, making it the best studied dye. This newly modified organic dye with enhanced PCE is remarkably effective for the dye-sensitized solar cells (DSSC) industry. Beyond materials discovery, this study highlights the role of high-performance computing in enabling predictive screening of dye candidates and generating performance indicators (HOMO-LUMO gaps, absorption spectra, charge transfer free energies, photovoltaic metrics). These outputs can serve as key parameters for energy informatics and system modelling.

Measurement of the Hall effect is a ubiquitous probe for materials discovery, characterization, and metrology. Inherent to the Hall measurement geometry, the measured signal is often contaminated by unwanted contributions, so the data must be processed to isolate the Hall response. The standard approach invokes Onsager-Casimir reciprocity and antisymmetrizes the raw signal about zero applied magnetic field. In hysteretic materials this becomes nontrivial, since Onsager-Casimir relations apply only to microscopically reversible states. Incorrect antisymmetrization can lead to artifacts that mimic anomalous or topological Hall signatures. The situation is especially subtle when hysteresis loops are not centered at zero applied field, as in exchange-biased systems. A practical reference for generically extracting the Hall response in hysteretic materials is lacking. Here, using Co$_3$Sn$_2$S$_2$ as a bulk single-crystal model that can be prepared with or without exchange-biased hysteresis, we demonstrate two procedures that can be used to extract the Hall effect: (1) reverse-magnetic-field reciprocity and (2) antisymmetrization with respect to applied field. We then measure the Hall effect on CeCoGe$_3$, a noncentrosymmetric antiferromagnet which can be prepared to have asymmetric magnetization and magnetoresistance, and demonstrate how improper processing can generate artificial anomalous Hall signals. These methods are generic and can be applied to any conductor.

We present a lattice-renormalized formalism for configurational tunneling two-level systems (TLS) that overcomes limitations of minimum-energy-path and light-particle models. Derived from the nuclear Hamiltonian, our formulation introduces composite phonon coordinates to capture lattice distortions between degenerate potential wells. This approach resolves deficiencies in prior models and enables accurate computation of tunnel splittings and excitation spectra for hydrogen-based TLS in bcc Nb. Our results bound experimental tunnel splittings and reveal strong anharmonic couplings between tunneling atoms and lattice phonons, establishing a direct link between TLS dynamics and phonon-mediated strain interactions. The formalism further generalizes to multi-level systems (MLS), providing insight into defect-induced decoherence in superconducting qubits and guiding strategies for materials design to suppress TLS-related loss.

We present a lattice-renormalized formalism for configurational tunneling two-level systems (TLS) that overcomes limitations of minimum-energy-path and light-particle models. Derived from the nuclear Hamiltonian, our formulation introduces composite phonon coordinates to capture lattice distortions between degenerate potential wells. This approach resolves deficiencies in prior models and enables accurate computation of tunnel splittings and excitation spectra for hydrogen-based TLS in bcc Nb. Our results bound experimental tunnel splittings and reveal strong anharmonic couplings between tunneling atoms and lattice phonons, establishing a direct link between TLS dynamics and phonon-mediated strain interactions. The formalism further generalizes to multi-level systems (MLS), providing insight into defect-induced decoherence in superconducting qubits and guiding strategies for materials design to suppress TLS-related loss.

Extracting structured information from zeolite synthesis experimental procedures is critical for materials discovery, yet existing methods have not systematically evaluated Large Language Models (LLMs) for this domain-specific task. This work addresses a fundamental question: what is the efficacy of different prompting strategies when applying LLMs to scientific information extraction? We focus on four key subtasks: event type classification (identifying synthesis steps), trigger text identification (locating event mentions), argument role extraction (recognizing parameter types), and argument text extraction (extracting parameter values). We evaluate four prompting strategies - zero-shot, few-shot, event-specific, and reflection-based - across six state-of-the-art LLMs (Gemma-3-12b-it, GPT-5-mini, O4-mini, Claude-Haiku-3.5, DeepSeek reasoning and non-reasoning) using the ZSEE dataset of 1,530 annotated sentences. Results demonstrate strong performance on event type classification (80-90\% F1) but modest performance on fine-grained extraction tasks, particularly argument role and argument text extraction (50-65\% F1). GPT-5-mini exhibits extreme prompt sensitivity with 11-79\% F1 variation. Notably, advanced prompting strategies provide minimal improvements over zero-shot approaches, revealing fundamental architectural limitations. Error analysis identifies systematic hallucination, over-generalization, and inability to capture synthesis-specific nuances. Our findings demonstrate that while LLMs achieve high-level understanding, precise extraction of experimental parameters requires domain-adapted models, providing quantitative benchmarks for scientific information extraction.

Advancing artificial intelligence for physical sciences requires representations that are both interpretable and compatible with the underlying laws of nature. We introduce METASTRINGS, a symbolic language for photonics that expresses nanostructures as textual sequences encoding materials, geometries, and lattice configurations. Analogous to molecular textual representations in chemistry, METASTRINGS provides a framework connecting human interpretability with computational design by capturing the structural hierarchy of photonic metasurfaces. Building on this representation, we develop Meta-GPT, a foundation transformer model trained on METASTRINGS and finetuned with physics-informed supervised, reinforcement, and chain-of-thought learning. Across various design tasks, the model achieves <3% mean-squared spectral error and maintains >98% syntactic validity, generating diverse metasurface prototypes whose experimentally measured optical responses match their target spectra. These results demonstrate that Meta-GPT can learn the compositional rules of light-matter interactions through METASTRINGS, laying a rigorous foundation for AI-driven photonics and representing an important step toward a metasurface genome project.

Conventional generative models for materials discovery are predominantly trained and validated using data from Density Functional Theory (DFT) with approximate exchange-correlation functionals. This creates a fundamental bottleneck: these models inherit DFT's systematic failures for strongly correlated systems, leading to exploration biases and an inability to discover materials where DFT predictions are qualitatively incorrect. We introduce a quantum-aware generative AI framework that systematically addresses this limitation through tight integration of multi-fidelity learning and active validation. Our approach employs a diffusion-based generator conditioned on quantum-mechanical descriptors and a validator using an equivariant neural network potential trained on a hierarchical dataset spanning multiple levels of theory (PBE, SCAN, HSE06, CCSD(T)). Crucially, we implement a robust active learning loop that quantifies and targets the divergence between low- and high-fidelity predictions. We conduct comprehensive ablation studies to deconstruct the contribution of each component, perform detailed failure mode analysis, and benchmark our framework against state-of-the-art generative models (CDVAE, GNoME, DiffCSP) across several challenging material classes. Our results demonstrate significant practical gains: a 3-5x improvement in successfully identifying potentially stable candidates in high-divergence regions (e.g., correlated oxides) compared to DFT-only baselines, while maintaining computational feasibility. This work provides a rigorous, transparent framework for extending the effective search space of computational materials discovery beyond the limitations of single-fidelity models.

Accelerating the discovery of structural materials is essential for applications in hard and refractory alloys, hypersonic platforms, nuclear systems, and other extreme environment technologies. Progress is often constrained by slow synthesis and characterization cycles and the need for extensive mechanical testing across large compositional spaces. Here, we propose a rapid screening strategy based on random material libraries, in which thousands of distinct compositions are embedded within a single specimen, mapped by EDS, and subsequently characterized. Using nanoindentation as a representative case, we show that such libraries enable dense composition property mapping while reducing the number of samples required to explore high dimensional composition spaces compared to traditional synthesis and test workflows. An experimentally calibrated Monte Carlo framework is developed to quantify practical limits, including particle size, EDS noise and resolution, positional accuracy, and nanoindenter motion costs. The simulations identify regimes where random libraries provide orders of magnitude acceleration over classical workflows. Finally, we demonstrate experimental navigation of these libraries using automated indentation. Together, these results establish random libraries as a general route to high throughput characterization in structurally critical material systems.

Large language models (LLMs) offer new opportunities for automated data extraction and property prediction across materials science, yet their use in superconductivity research remains limited. Here we construct a large experimental database of 78,203 records, covering 19,058 unique compositions, extracted from scientific literature using an LLM-driven workflow. Each entry includes chemical composition, critical temperature, measurement pressure, structural descriptors, and critical fields. We fine-tune several open-source LLMs for three tasks: (i) classifying superconductors vs. non-superconductors, (ii) predicting the superconducting transition temperature directly from composition or structure-informed inputs, and (iii) inverse design of candidate compositions conditioned on target Tc. The fine-tuned LLMs achieve performance comparable to traditional feature-based models and in some cases exceed them, while substantially outperforming their base versions and capturing meaningful chemical and structural trends. The inverse-design model generates chemically plausible compositions, including 28% novel candidates not seen in training. Finally, applying the trained predictors to the GNoME database identifies unreported materials with predicted Tc > 10 K. Although unverified, these candidates illustrate how integrating an LLM-driven workflow can enable scalable hypothesis generation for superconductivity discovery.

Inverse design of heterogeneous material microstructures is a fundamentally ill-posed and famously computationally expensive problem. This is exacerbated by the high-dimensional design spaces associated with finely resolved images, multimodal input property streams, and a highly nonlinear forward physics. Whilst modern generative models excel at accurately modeling such complex forward behavior, most of them are not intrinsically structured to support fast, stable \emph{deterministic} inversion with a physics-informed bias. This work introduces Janus, a unified generative-predictive framework to address this problem. Janus couples a deep encoder-decoder architecture with a predictive KHRONOS head, a separable neural architecture. Topologically speaking, Janus learns a latent manifold simultaneously isometric for generative inversion and pruned for physical prediction; the joint objective inducing \emph{disentanglement} of the latent space. Janus is first validated on the MNIST dataset, demonstrating high-fidelity reconstruction, accurate classification and diverse generative inversion of all ten target classes. It is then applied to the inverse design of heterogeneous microstructures labeled with thermal conductivity. It achieves a forward prediction accuracy $R^2=0.98$ (2\% relative error) and sub-5\% pixelwise reconstruction error. Inverse solutions satisfy target properties to within $1\%$ relative error. Inverting a sweep through properties reveal smooth traversal of the latent manifold, and UMAP visualization confirms the emergence of a low-dimensional, disentangled manifold. By unifying prediction and generation within a single latent space, Janus enables real-time, physics-informed inverse microstructure generation at a lower computational cost typically associated with classical optimization-based approaches.

Recent advances in generative machine learning have opened new possibilities for the discovery and design of novel materials. However, as these models become more sophisticated, the need for rigorous and meaningful evaluation metrics has grown. Existing evaluation approaches often fail to capture both the quality and novelty of generated structures, limiting our ability to assess true generative performance. In this paper, we introduce the Transport Novelty Distance (TNovD) to judge generative models used for materials discovery jointly by the quality and novelty of the generated materials. Based on ideas from Optimal Transport theory, TNovD uses a coupling between the features of the training and generated sets, which is refined into a quality and memorization regime by a threshold. The features are generated from crystal structures using a graph neural network that is trained to distinguish between materials, their augmented counterparts, and differently sized supercells using contrastive learning. We evaluate our proposed metric on typical toy experiments relevant for crystal structure prediction, including memorization, noise injection and lattice deformations. Additionally, we validate the TNovD on the MP20 validation set and the WBM substitution dataset, demonstrating that it is capable of detecting both memorized and low-quality material data. We also benchmark the performance of several popular material generative models. While introduced for materials, our TNovD framework is domain-agnostic and can be adapted for other areas, such as images and molecules.

We present a novel multi-stage workflow for computational materials discovery that achieves a 99% success rate in identifying compounds within 100 meV/atom of thermodynamic stability, with a threefold improvement over previous approaches. By combining the Matra-Genoa generative model, Orb-v2 universal machine learning interatomic potential, and ALIGNN graph neural network for energy prediction, we generated 119 million candidate structures and added 1.3 million DFT-validated compounds to the ALEXANDRIA database, including 74 thousand new stable materials. The expanded ALEXANDRIA database now contains 5.8 million structures with 175 thousand compounds on the convex hull. Predicted structural disorder rates (37-43%) match experimental databases, unlike other recent AI-generated datasets. Analysis reveals fundamental patterns in space group distributions, coordination environments, and phase stability networks, including sub-linear scaling of convex hull connectivity. We release the complete dataset, including sAlex25 with 14 million out-of-equilibrium structures containing forces and stresses for training universal force fields. We demonstrate that fine-tuning a GRACE model on this data improves benchmark accuracy. All data, models, and workflows are freely available under Creative Commons licenses.

Machine learning (ML) offers a powerful path toward discovering sustainable polymer materials, but progress has been limited by the lack of large, high-quality, and openly accessible polymer datasets. The Open Polymer Challenge (OPC) addresses this gap by releasing the first community-developed benchmark for polymer informatics, featuring a dataset with 10K polymers and 5 properties: thermal conductivity, radius of gyration, density, fractional free volume, and glass transition temperature. The challenge centers on multi-task polymer property prediction, a core step in virtual screening pipelines for materials discovery. Participants developed models under realistic constraints that include small data, label imbalance, and heterogeneous simulation sources, using techniques such as feature-based augmentation, transfer learning, self-supervised pretraining, and targeted ensemble strategies. The competition also revealed important lessons about data preparation, distribution shifts, and cross-group simulation consistency, informing best practices for future large-scale polymer datasets. The resulting models, analysis, and released data create a new foundation for molecular AI in polymer science and are expected to accelerate the development of sustainable and energy-efficient materials. Along with the competition, we release the test dataset at https://www.kaggle.com/datasets/alexliu99/neurips-open-polymer-prediction-2025-test-data. We also release the data generation pipeline at https://github.com/sobinalosious/ADEPT, which simulates more than 25 properties, including thermal conductivity, radius of gyration, and density.

Since the advent of machine learning, interpretability has remained a persistent challenge, becoming increasingly urgent as generative models support high-stakes applications in drug and material discovery. Recent advances in large language model (LLM) architectures have yielded chemistry language models (CLMs) with impressive capabilities in molecular property prediction and molecular generation. However, how these models internally represent chemical knowledge remains poorly understood. In this work, we extend sparse autoencoder techniques to uncover and examine interpretable features within CLMs. Applying our methodology to the Foundation Models for Materials (FM4M) SMI-TED chemistry foundation model, we extract semantically meaningful latent features and analyse their activation patterns across diverse molecular datasets. Our findings reveal that these models encode a rich landscape of chemical concepts. We identify correlations between specific latent features and distinct domains of chemical knowledge, including structural motifs, physicochemical properties, and pharmacological drug classes. Our approach provides a generalisable framework for uncovering latent knowledge in chemistry-focused AI systems. This work has implications for both foundational understanding and practical deployment; with the potential to accelerate computational chemistry research.

High-performance concrete requires complex mix design decisions involving interdependent variables and practical constraints. While data-driven methods have improved predictive modeling for forward design in concrete engineering, inverse design remains limited, especially when some variables are fixed and only the remaining ones must be inferred. This study proposes a cooperative neural network framework for the partial inverse design of high-performance concrete. The framework integrates an imputation model with a surrogate strength predictor and learns through cooperative training. Once trained, it generates valid and performance-consistent mix designs in a single forward pass without retraining for different constraint scenarios. Compared with baseline models, including autoencoder models and Bayesian inference with Gaussian process surrogates, the proposed method achieves R-squared values of 0.87 to 0.92 and substantially reduces mean squared error by approximately 50% and 70%, respectively. The results show that the framework provides an accurate and computationally efficient foundation for constraint-aware, data-driven mix proportioning.

High-performance concrete offers exceptional strength and durability but requires complex mix designs involving many interdependent variables and practical constraints. While data-driven methods have advanced predictive modeling for forward design, inverse design, which focuses on determining mix compositions that achieve target performance, remains limited, particularly in design situations where some mix variables are fixed by constraints and only the remaining variables must be determined. This study proposes a cooperative neural network framework for the partial inverse design of high-performance concrete. The framework combines two coupled neural network models, an imputation model that infers the undetermined variables and a surrogate model that predicts compressive strength. Through cooperative learning, the model generates valid and performance-consistent mix designs in a single forward pass while accommodating different constraint combinations without retraining. Its performance is compared with both probabilistic and generative approaches, including Bayesian inference based on a Gaussian process surrogate and autoencoder-based models. Evaluated on a benchmark dataset, the proposed model achieves stable and higher R-squared values of 0.87-0.92 and reduces mean squared error by an average of 50 percent compared with autoencoder baselines and by an average of 70 percent compared with Bayesian inference. The results demonstrate that the cooperative neural network provides an accurate, robust, and computationally efficient foundation for constraint-aware, data-driven mix proportioning in concrete engineering.

Inverse design aims to design the input variables of a physical system to optimize a specified objective function, typically formulated as a search or optimization problem. However, in 3D domains, the design space grows exponentially, rendering exhaustive grid-based searches infeasible. Recent advances in deep learning have accelerated inverse design by providing powerful generative priors and differentiable surrogate models. Nevertheless, current methods tend to approximate the 3D design space using 2D projections or fine-tune existing 3D shapes. These approaches sacrifice volumetric detail and constrain design exploration, preventing true 3D design from scratch. In this paper, we propose a 3D Inverse Design (3DID) framework that directly navigates the 3D design space by coupling a continuous latent representation with a physics-aware optimization strategy. We first learn a unified physics-geometry embedding that compactly captures shape and physical field data in a continuous latent space. Then, we introduce a two-stage strategy to perform physics-aware optimization. In the first stage, a gradient-guided diffusion sampler explores the global latent manifold. In the second stage, an objective-driven, topology-preserving refinement further sculpts each candidate toward the target objective. This enables 3DID to generate high-fidelity 3D geometries, outperforming existing methods in both solution quality and design versatility.

Predictive atomistic simulations have propelled materials discovery, yet routine setup and debugging still demand computer specialists. This know-how gap limits Integrated Computational Materials Engineering (ICME), where state-of-the-art codes exist but remain cumbersome for non-experts. We address this bottleneck with GENIUS, an AI-agentic workflow that fuses a smart Quantum ESPRESSO knowledge graph with a tiered hierarchy of large language models supervised by a finite-state error-recovery machine. Here we show that GENIUS translates free-form human-generated prompts into validated input files that run to completion on $\approx$80% of 295 diverse benchmarks, where 76% are autonomously repaired, with success decaying exponentially to a 7% baseline. Compared with LLM-only baselines, GENIUS halves inference costs and virtually eliminates hallucinations. The framework democratizes electronic-structure DFT simulations by intelligently automating protocol generation, validation, and repair, opening large-scale screening and accelerating ICME design loops across academia and industry worldwide.

Machine learning has transformed material discovery for inorganic compounds and small molecules, yet polymers remain largely inaccessible to these methods. While data scarcity is often cited as the primary bottleneck, we demonstrate that strategic molecular representations can overcome this limitation. We introduce CI-LLM (Chemically Informed Language Model), a framework combining HAPPY (Hierarchically Abstracted rePeat unit of PolYmer), which encodes chemical substructures as tokens, with numerical descriptors within transformer architectures. For property prediction, De$^3$BERTa, our descriptor-enriched encoder, achieves 3.5x faster inference than SMILES-based models with improved accuracy ($R^2$ score gains of 0.9-4.1 percent across four properties), while providing interpretable structure-property insights at the subgroup level. For inverse design, our GPT-based generator produces polymers with targeted properties, achieving 100 percent scaffold retention and successful multi-property optimization for negatively correlated objectives. This comprehensive framework demonstrates both forward prediction and inverse design capabilities, showcasing how strategic molecular representation advances machine learning applications in polymer science.

The observation of reversible de-mixing phenomena in mixed-halide perovskites under illumination is one of the most challenging as well as intriguing aspects of this class of materials. On the one hand, it poses critical constraints to the compositional space that allows reliable design of absorbers for perovskite photovoltaics. On the other hand, it holds potential for the development of novel optoionic devices where an ionic response is triggered via optical stimuli. Funda-mental questions about the origin of such photo de-mixing process remain unanswered, both in terms of its mechanism as well as thermodynamic description. Here, we relate in-situ measurements of ionic and electronic transport of mixed bromide-iodide perovskite thin films performed during photo de-mixing with the evolution of their optical and morpho-logical properties. The results point to the definition of different stages of the de-mixing process which, based on micros-copy and spectroscopic measurements, we assign to regimes of spinodal decomposition and nucleation of quasi-equilibrium iodide- and bromide-rich phases. Combined with density functional theory calculations, we explore the role of dimensionality in the mechanism and reversibility of photo de-mixing and dark re-mixing processes, referring to elec-tronic and ionic contributions to the de-mixing driving force. Additionally, our data emphasizes the role of the surface, as significantly different de-mixing dynamics, in terms of extent and reversibility, are observed for films with or without encapsulation. Our comprehensive analysis of transport, phase and optical properties of mixed-halide perovskites pro-vides guidelines for future materials design as well as for the more general fundamental understanding of light-induced ionic phenomena.

We introduce ECHO, a transformer-operator framework for generating million-point PDE trajectories. While existing neural operators (NOs) have shown promise for solving partial differential equations, they remain limited in practice due to poor scalability on dense grids, error accumulation during dynamic unrolling, and task-specific design. ECHO addresses these challenges through three key innovations. (i) It employs a hierarchical convolutional encode-decode architecture that achieves a 100 $\times$ spatio-temporal compression while preserving fidelity on mesh points. (ii) It incorporates a training and adaptation strategy that enables high-resolution PDE solution generation from sparse input grids. (iii) It adopts a generative modeling paradigm that learns complete trajectory segments, mitigating long-horizon error drift. The training strategy decouples representation learning from downstream task supervision, allowing the model to tackle multiple tasks such as trajectory generation, forward and inverse problems, and interpolation. The generative model further supports both conditional and unconditional generation. We demonstrate state-of-the-art performance on million-point simulations across diverse PDE systems featuring complex geometries, high-frequency dynamics, and long-term horizons.

An Ag-doped CuO-AgCl-g-C3N4 heterostructure has been designed to achieve rapid Methylene Blue (MB) degradation through a synergistic photo-Fenton mechanism driven by low-power UV illumination. The composite integrates narrow-bandgap CuO, plasmonic Ag/AgCl, and visible-responsive g-C3N4 into a dual Z-scheme configuration that promotes efficient interfacial charge transfer while preserving strong redox potentials. Diffuse reflectance UV-Vis spectra ascertained the bandgap positions of the composite corresponding to those of its constituents: 2.9 eV (g-C3N4) and 1.42 eV (Ag-doped CuO-AgCl), indicating enhanced absorption and efficient charge carrier generation. BET analysis confirmed the presence of mesoporosity and revealed an effective surface area, ensuring the availability of abundant adsorption and reaction sites. A commercial 11 W UV irradiation was used for the photocatalytic test. Almost complete degradation of MB occurred within 10 min, following pseudo-first-order kinetics with a high apparent rate constant of 0.45/min. The remarkable activity arises from the synergistic interplay of Fenton-like redox cycling and efficient photoinduced charge carrier generation and separation. In addition, it has been demonstrated that intentionally incorporated AgCl plays an active role as a plasmonic-semiconducting interface, strengthening charge separation and catalyst stability under neutral conditions, rather than acting as a passive chloride byproduct. Overall, by linking defect engineering, heterojunction design, and photo-Fenton synergy, this study establishes a low-power, catalytic platform offering a viable pathway towards sustainable dye wastewater remediation.

We present a first-principles study of the nonlinear optical properties of sliding ferroelectric bilayers based on MoSe$_2$ and Janus MoSSe. Two Janus configurations are considered: i) one bilayer where the two intralayer polarizations caused by Janus chemical asymmetry cancel each other out, yielding photocurrent spectra comparable to pristine MoSe$_2$ bilayers; ii) another bilayer where the intralayer polarizations add up, for which the photoresponses are strongly enhanced. Our results show that photocurrent generation in the polar Janus structures is predominantly governed by vertical chemical asymmetry, with limited dependence on the sliding direction. These findings highlight complementary design strategies: interlayer sliding enables sensitivity to external tuning, while the Janus intralayer polarization enhances photoresponses in the visible range. The interplay between composition and stacking therefore provides a versatile platform for tailoring light-matter interactions in 2D ferroelectric materials.

Discovering nonlinear optical (NLO) materials with strong shift current response, particularly in the infrared (IR) regime, is essential for next-generation optoelectronics yet remains highly challenging in both experiments and theory, which still largely relies on case by case studies. Here, we employ a high-throughput screening strategy, applying a multi-step filter to the Materials Project database (>154,000 materials), which yielded 2,519 candidate materials for detailed first-principle evaluation. From these calculations, we identify 32 NLO materials with strong shift current response ($σ$ > 100 $μA/V^2$). Our work reveals that layered structures with $C_{3v}$ symmetry and heavy $p$-block elements (e.g. Te, Sb) exhibit apparent superiority in enhancing shift current. More importantly, 9 of these compounds show shift current response peaks in the IR region, with the strongest reaching 616 $μA/V^2$, holding significant application potential in fields such as IR photodetection, sensing, and energy harvesting. Beyond identifying promising candidates, this work establishes a comprehensive and high-quality first-principles dataset for NLO response, providing a solid foundation for future AI-driven screening and accelerated discovery of high-performance NLO materials, as demonstrated by a prototype machine-learning application.

Generative machine learning (ML) models hold great promise for accelerating materials discovery through the inverse design of inorganic crystals, enabling an unprecedented exploration of chemical space. Yet, the lack of standardized evaluation frameworks makes it challenging to evaluate, compare, and further develop these ML models meaningfully. In this work, we introduce LeMat-GenBench, a unified benchmark for generative models of crystalline materials, supported by a set of evaluation metrics designed to better inform model development and downstream applications. We release both an open-source evaluation suite and a public leaderboard on Hugging Face, and benchmark 12 recent generative models. Results reveal that an increase in stability leads to a decrease in novelty and diversity on average, with no model excelling across all dimensions. Altogether, LeMat-GenBench establishes a reproducible and extensible foundation for fair model comparison and aims to guide the development of more reliable, discovery-oriented generative models for crystalline materials.

Generative machine learning (ML) models hold great promise for accelerating materials discovery through the inverse design of inorganic crystals, enabling an unprecedented exploration of chemical space. Yet, the lack of standardized evaluation frameworks makes it challenging to evaluate, compare, and further develop these ML models meaningfully. In this work, we introduce LeMat-GenBench, a unified benchmark for generative models of crystalline materials, supported by a set of evaluation metrics designed to better inform model development and downstream applications. We release both an open-source evaluation suite and a public leaderboard on Hugging Face, and benchmark 12 recent generative models. Results reveal that an increase in stability leads to a decrease in novelty and diversity on average, with no model excelling across all dimensions. Altogether, LeMat-GenBench establishes a reproducible and extensible foundation for fair model comparison and aims to guide the development of more reliable, discovery-oriented generative models for crystalline materials.

We present a data-efficient framework for constructing general classical spin Hamiltonians from the spin-cluster expansion (SCE) combined with fully self-consistent noncollinear spin density functional theory (DFT). The key idea is to fit an SCE model to magnetic torques rather than to total energies. Because torques are site-resolved vectors, each configuration supplies many independent constraints, which makes the regression well conditioned and sharply reduces the number of DFT calculations needed, especially in large supercells. Applied to the B20-type chiral magnets ${\rm Mn}_{1-x}{\rm Fe}_{x}{\rm Ge}$ and ${\rm Fe}_{1-y}{\rm Co}_{y}{\rm Ge}$, the resulting models nonperturbatively extract the full pairwise exchange tensor (isotropic exchange, anisotropic symmetric exchange, and the Dzyaloshinskii--Moriya interaction) and predict helical spin period via a micromagnetic mapping. The composition trends and the divergence of the period near the chirality sign change are reproduced in line with experiments. Because the SCE framework is systematic, it also enables systematic analysis of interaction order; training on increasingly disordered spin configurations shows that the lowest-order model loses torque accuracy, whereas including higher-order interactions restores predictive power. These advances enable near-DFT-accurate spin models for finite-temperature magnetism and complex textures at modest data cost, while providing a systematic, extensible, and nonperturbative route to quantitative first-principles parameterization and predictive materials design. An open-source implementation is available as the Julia package, \textit{Magesty.jl}.

We present a data-efficient framework for constructing general classical spin Hamiltonians by combining the spin-cluster expansion (SCE) with fully self-consistent noncollinear spin density functional theory (DFT). The key idea is to fit the SCE model to magnetic torques rather than to total energies. Because torques are site-resolved vectors, each spin configuration provides many informative regression targets, improving conditioning and substantially reducing the number of required DFT calculations, especially for large supercells. Applied to the B20-type chiral magnets ${\rm Mn}_{1-x}{\rm Fe}_{x}{\rm Ge}$ and ${\rm Fe}_{1-y}{\rm Co}_{y}{\rm Ge}$, the resulting SCE models determine full pairwise exchange tensors -- including isotropic exchange, symmetric anisotropic exchange, and the Dzyaloshinskii--Moriya interaction -- and predict the helical spin period via a micromagnetic mapping. The composition trends and the divergence of the period at the chirality sign-change point are well reproduced, in agreement with experiment. Moreover, the systematic nature of SCE enables controlled assessment of interaction order: as the training spin configurations become more disordered, the lowest-order model loses torque accuracy, whereas including higher-order interactions restores predictive power. These advances enable near-DFT-accurate spin models for finite-temperature magnetism and complex spin textures at modest computational cost, providing an extensible route to quantitative first-principles parameterization and predictive materials design. An open-source implementation is available as a Julia package, \textit{Magesty.jl}.

High-resolution transmission electron microscopy (HRTEM) is crucial for observing material's structural and morphological evolution at Angstrom scales, but the electron beam can alter these processes. Devices such as CMOS-based direct-electron detectors operating in electron-counting mode can be utilized to substantially reduce the electron dosage. However, the resulting images often lead to low signal-to-noise ratio, which requires frame integration that sacrifices temporal resolution. Several machine learning (ML) models have been recently developed to successfully denoise HRTEM images. Yet, these models are often computationally expensive and their inference speeds on GPUs are outpaced by the imaging speed of advanced detectors, precluding in situ analysis. Furthermore, the performance of these denoising models on datasets with imaging conditions that deviate from the training datasets have not been evaluated. To mitigate these gaps, we propose a new self-supervised ML denoising pipeline specifically designed for time-series HRTEM images. This pipeline integrates a blind-spot convolution neural network with pre-processing and post-processing steps including drift correction and low-pass filtering. Results demonstrate that our model outperforms various other ML and non-ML denoising methods in noise reduction and contrast enhancement, leading to improved visual clarity of atomic features. Additionally, the model is drastically faster than U-Net-based ML models and demonstrates excellent out-of-distribution generalization. The model's computational inference speed is in the order of milliseconds per image, rendering it suitable for application in in-situ HRTEM experiments.

This thesis presents novel contributions in two primary areas: advancing the efficiency of generative models, particularly normalizing flows, and applying generative models to solve real-world computer vision challenges. The first part introduce significant improvements to normalizing flow architectures through six key innovations: 1) Development of invertible 3x3 Convolution layers with mathematically proven necessary and sufficient conditions for invertibility, (2) introduction of a more efficient Quad-coupling layer, 3) Design of a fast and efficient parallel inversion algorithm for kxk convolutional layers, 4) Fast & efficient backpropagation algorithm for inverse of convolution, 5) Using inverse of convolution, in Inverse-Flow, for the forward pass and training it using proposed backpropagation algorithm, and 6) Affine-StableSR, a compact and efficient super-resolution model that leverages pre-trained weights and Normalizing Flow layers to reduce parameter count while maintaining performance. The second part: 1) An automated quality assessment system for agricultural produce using Conditional GANs to address class imbalance, data scarcity and annotation challenges, achieving good accuracy in seed purity testing; 2) An unsupervised geological mapping framework utilizing stacked autoencoders for dimensionality reduction, showing improved feature extraction compared to conventional methods; 3) We proposed a privacy preserving method for autonomous driving datasets using on face detection and image inpainting; 4) Utilizing Stable Diffusion based image inpainting for replacing the detected face and license plate to advancing privacy-preserving techniques and ethical considerations in the field.; and 5) An adapted diffusion model for art restoration that effectively handles multiple types of degradation through unified fine-tuning.

A topological interlocking assembly is an arrangement of blocks, where all blocks are kinematically constrained by their neighboring blocks and a fixed frame. This concept has been known for a long time, attracting recent interest due to its advantageous mechanical properties, such as reusability, redundancy and limited crack propagation. New mathematical methods enable the generation of vast numbers of new topologically interlocking blocks. A natural next question is the quantification of the mechanical performance of these new blocks. We conduct a numerical study of topological interlocking assemblies whose blocks are constructed based on the hexagonal grid. By varying a design parameter used in the generation of these blocks, we study its influence on the structural performance of the entire assembly. The results improve our understanding of the link between the block parameters and the mechanical performance. This enhances the ability to custom design blocks for certain mechanical requirements of the topological interlocking assemblies.

Fast, and accurate prediction of ionic migration barriers ($E_m$) is crucial for designing next-generation battery materials that combine high energy density with facile ion transport. Given the computational costs associated with estimating $E_m$ using conventional density functional theory (DFT) based nudged elastic band (NEB) calculations, we benchmark the accuracy in $E_m$ and geometry predictions of five foundational machine learned interatomic potentials (MLIPs), which can potentially accelerate predictions of ionic transport. Specifically, we assess the accuracy of MACE-MP-0, Orb-v3, SevenNet, CHGNet, and M3GNet models, coupled with the NEB framework, against DFT-NEB-calculated $E_m$ across a diverse set of battery-relevant chemistries and structures. Notably, MACE-MP-0 and Orb-v3 exhibit the lowest mean absolute errors in $E_m$ predictions across the entire dataset and over data points that are not outliers, respectively. Importantly, Orb-v3 and SevenNet classify `good' versus `bad' ionic conductors with an accuracy of $>$82\%, based on a threshold $E_m$ of 500~meV, indicating their utility in high-throughput screening approaches. Notably, intermediate images generated by MACE-MP-0 and SevenNet provide better initial guesses relative to conventional interpolation techniques in $>$71\% of structures, offering a practical route to accelerate subsequent DFT-NEB relaxations. Finally, we observe that accurate $E_m$ predictions by MLIPs are not correlated with accurate (local) geometry predictions. Our work establishes the use-cases, accuracies, and limitations of foundational MLIPs in estimating $E_m$ and should serve as a base for accelerating the discovery of novel ionic conductors for batteries and beyond.

Electrical impedance tomography (EIT) provides an attractive solution for large-area tactile sensing due to its minimal wiring and shape flexibility, but its nonlinear inverse problem often leads to severe artifacts and inaccurate contact reconstruction. This work presents PhyDNN, a physics-driven deep reconstruction framework that embeds the EIT forward model directly into the learning objective. By jointly minimizing the discrepancy between predicted and ground-truth conductivity maps and enforcing consistency with the forward PDE, PhyDNN reduces the black-box nature of deep networks and improves both physical plausibility and generalization. To enable efficient backpropagation, we design a differentiable forward-operator network that accurately approximates the nonlinear EIT response, allowing fast physics-guided training. Extensive simulations and real tactile experiments on a 16-electrode soft sensor show that PhyDNN consistently outperforms NOSER, TV, and standard DNNs in reconstructing contact shape, location, and pressure distribution. PhyDNN yields fewer artifacts, sharper boundaries, and higher metric scores, demonstrating its effectiveness for high-quality tomographic tactile sensing.

The rapid advancement and adaptability of Large Language Models (LLMs) highlight the need for moral consistency, the capacity to maintain ethically coherent reasoning across varied contexts. Existing alignment frameworks, structured approaches designed to align model behavior with human ethical and social norms, often rely on static datasets and post-hoc evaluations, offering limited insight into how ethical reasoning may evolve across different contexts or temporal scales. This study presents the Moral Consistency Pipeline (MoCoP), a dataset-free, closed-loop framework for continuously evaluating and interpreting the moral stability of LLMs. MoCoP combines three supporting layers: (i) lexical integrity analysis, (ii) semantic risk estimation, and (iii) reasoning-based judgment modeling within a self-sustaining architecture that autonomously generates, evaluates, and refines ethical scenarios without external supervision. Our empirical results on GPT-4-Turbo and DeepSeek suggest that MoCoP effectively captures longitudinal ethical behavior, revealing a strong inverse relationship between ethical and toxicity dimensions (correlation rET = -0.81, p value less than 0.001) and a near-zero association with response latency (correlation rEL approximately equal to 0). These findings demonstrate that moral coherence and linguistic safety tend to emerge as stable and interpretable characteristics of model behavior rather than short-term fluctuations. Furthermore, by reframing ethical evaluation as a dynamic, model-agnostic form of moral introspection, MoCoP offers a reproducible foundation for scalable, continuous auditing and advances the study of computational morality in autonomous AI systems.

Spin field-effect transistors (SFETs) are promising candidates for low-power spin-based electronics, yet existing realizations that rely on spin-orbit coupling are constrained by limited material choices and short spin-coherence lengths. Here we propose a different operating principle based on multiferroic altermagnets, in which spin splitting is tuned by an electric field through symmetry control rather than conventional spin-orbit physics. Using an effective model combined with quantum transport simulations, we show that the conductance is determined by the degree of matching between the electrically controlled spin texture of the channel and the fixed spin polarization of ferromagnetic contacts, enabling clear ON and OFF states. Remarkably, we also address a long-standing challenge in multiferroic device design: spintronic channels require metallic carriers, whereas ferroelectricity is usually suppressed in metals. We resolve this conflict by imprinting multiferroic altermagnetism into highly conductive materials via the proximity effect. First-principles calculations for graphene on multiferroic vanadium sulfide halides confirm that graphene acquires a ferroelectrically switchable spin splitting while retaining its metallic character. These results establish a practical route to SFET implementation and identify multiferroic altermagnets as a versatile platform for next-generation spintronic devices.

Adaptive material systems that autonomously respond to external stimuli are crucial for advancing next-generation smart devices. Biological systems achieve autonomous behavior by utilizing chemical energy from out-of-equilibrium reactions to power life-like functions without requiring external energy inputs. Although responsive hydrogels with embedded enzymatic reactions offer a promising platform for implementing adaptive behavior in synthetic systems, previous studies have focused on controlling the supramolecular self-assembly of responsive building blocks rather than modulating network crosslinking. Here, we demonstrate direct enzymatic modulation of crosslinking density in a double-network hydrogel to achieve autonomous self-stiffening in response to a chemical stimulus. Our adaptive system embeds glucose oxidase within a polyacrylamide-alginate double-network hydrogel containing Ca(EDTA)2- complexes that render the crosslinked alginate network pH-responsive through a competitive calcium binding mechanism. Chemical waves emerging from enzymatic reaction activation propagate at speeds ranging from 15 to 44 um/min, driving spatiotemporal mechanical transitions that increase material stiffness by up to 2.1-fold. By integrating signal sensing and chemomechanical transduction within this responsive hydrogel, we realized adaptive behavior that autonomously converts localized chemical inputs into system-wide mechanical outputs. This positions our adaptive hydrogels as promising model systems to guide the design of intelligent materials for soft robotics and biomedical devices.

With the rapid development of the online education and large language model, the existing educational tools still suffer from incomplete service, insufficient performance and weak interactivity in terms of courseware generation, interactive notes and quality assurance of content. In particular, the proposed generative agent EZYer : 1) Teacher Module: Integrating the Text Corpus retrieval and in-depth generation technologies, it automatically generates structured teaching materials and LaTeX Beamer courseware in line with the high school mathematics syllabus and supports user-defined image insertion. 2) Student Module: Throughout the collaborative interaction of the four roles of Teacher, Assistant, Top Student and Struggling Student, Note Taker summarizes and generates academic notes to enhance the depth and interest of learning. 3) Controller: set up keyword filtering system, content scoring system, role co-validation system, and dynamic content correction system. This ensure academic strictness and pedagogical propriety of EZYer inputs and outputs. In order to evaluate EZYer, this paper designs five-dimensional evaluation indexes of content accuracy, knowledge coverage, usability, formatting correctness and visual design and appeal, and scores 100 Beamer and Notes generated by EZYer by five large language models, separately, and the results show that the quality of EZYer-generated content is excellent and has a good application prospect.

Perovskite ruthenates are fascinating playgrounds for exploring topological spin textures, but generally rely on extrinsic mechanisms to trigger the noncoplanar states. Here we report the discovery of an emergent chiral spin crystal phase in (111) SrRuO3 epitaxial films, characterized by a significant topological Hall effect and noncoplanar spin arrangements with different propagation vectors along two orthogonal directions. Instead of driven by the enhanced Dzyaloshinskii-Moriya interaction due to broken inversion symmetry at heterointerfaces, this emergent state arises intrinsically from the interplay of dipolar interactions and magnetic frustration, leading to the stabilization of topological phases in much thicker films. These findings open a new pathway for creating and controlling the topological spin states in perovskites, with broad implications for spintronic device design.

Complex concentrated alloys with intrinsic chemical heterogeneity are promising candidates for nuclear applications, where local chemical order can strongly influence defect evolution under irradiation. Grain boundaries also contribute to radiation damage mitigation by serving as defect sinks, yet this interaction can alter interfacial structure, typically leading to destabilization and grain growth. This study investigates how chemical ordering influences grain boundary migration and stability during successive radiation events in CrCoNi. Using atomistic simulations, bicrystals were equilibrated to induce segregation-enhanced chemical order, followed by prolonged irradiation at 1100 K. Our results show that grain boundaries in random CrCoNi begin to migrate after only a few collision cascades, whereas those in the ordered alloy remain immobile until the chemical order is sufficiently disrupted. Single-cascade simulations reveal key mechanistic differences, where cascades near chemically ordered interfaces produce smaller damage volumes and reduced atomic displacement due to enhanced Frenkel pair combination within the cascade core. This limits both the residual defect population and the energetic driving force for interfacial rearrangement. Subsequent simulations of irradiated interfaces show that interstitial absorption induces a structural transition that modifies the segregation morphology at and near the grain boundary, demonstrating a dynamic coupling between ordering stability and defect evolution. These findings offer new insights into the role of local chemical order on defect-interface interactions under extreme conditions and highlight pathways for designing radiation-tolerant materials for next-generation nuclear systems.

Boron aluminum nitride (B$_x$Al$_{1-x}$N) is a promising material for next-generation electronic and optoelectronic devices due to its ultra-wide bandgap, high thermal stability, and compatibility with other III-nitride semiconductors. Despite its potential, the band alignments of B$_x$Al$_{1-x}$N remain largely unexplored, although this information is essential for device design. In this study, we compute the valence and conduction band alignments of nonpolar ($a$-plane) and polar ($c$-plane) B$_x$Al$_{1-x}$N, and compare them with those of AlN and GaN. Using density functional theory, many-body perturbation theory, $GW_0$ method, and a novel passivation scheme, we find that they have near-zero valence band alignments for low-$x$ B$_x$Al$_{1-x}$N/AlN, while higher compositions ($x > $0.333) exhibit type I or II band alignments. The band alignments also show a notable dependence on surface polarity and the tetrahedral distortion of the B$_x$Al$_{1-x}$N structures. Our computed offsets are in good agreement with available experimental data. Due to their low valence band alignments and higher conduction band alignments, the B$_x$Al$_{1-x}$N/AlN heterostructures could be well suited for high-electron-mobility transistors and ultraviolet light-emitting diodes. The band alignments of B$_x$Al$_{1-x}$N determined in this study provide essential design guidelines for integrating these ultra-wide bandgap alloys into advanced semiconductor technologies.

Mofasa is an all-atom latent diffusion model with state-of-the-art performance for generating Metal-Organic Frameworks (MOFs). These are highly porous crystalline materials used to harvest water from desert air, capture carbon dioxide, store toxic gases and catalyse chemical reactions. In recognition of their value, the development of MOFs recently received a Nobel Prize in Chemistry. In many ways, MOFs are well-suited for exploiting generative models in chemistry: they are rationally-designable materials with a large combinatorial design space and strong structure-property couplings. And yet, to date, a high performance generative model has been lacking. To fill this gap, we introduce Mofasa, a general-purpose latent diffusion model that jointly samples positions, atom-types and lattice vectors for systems as large as 500 atoms. Mofasa avoids handcrafted assembly algorithms common in the literature, unlocking the simultaneous discovery of metal nodes, linkers and topologies. To help the scientific community build on our work, we release MofasaDB, an annotated library of hundreds of thousands of sampled MOF structures, along with a user-friendly web interface for search and discovery: https://mofux.ai/ .

Large language model (LLM)-based agentic frameworks increasingly adopt the paradigm of dynamically generating task-specific agents. We suggest that not only agents but also specialized software modules for scientific and engineering tasks can be generated on demand. We demonstrate this concept in the field of solid mechanics. There, so-called constitutive models are required to describe the relationship between mechanical stress and body deformation. Constitutive models are essential for both the scientific understanding and industrial application of materials. However, even recent data-driven methods of constitutive modeling, such as constitutive artificial neural networks (CANNs), still require substantial expert knowledge and human labor. We present a framework in which an LLM generates a CANN on demand, tailored to a given material class and dataset provided by the user. The framework covers LLM-based architecture selection, integration of physical constraints, and complete code generation. Evaluation on three benchmark problems demonstrates that LLM-generated CANNs achieve accuracy comparable to or greater than manually engineered counterparts, while also exhibiting reliable generalization to unseen loading scenarios and extrapolation to large deformations. These findings indicate that LLM-based generation of physics-constrained neural networks can substantially reduce the expertise required for constitutive modeling and represent a step toward practical end-to-end automation.

Digital Twins (DTs) are increasingly used as autonomous decision-makers in complex socio-technical systems. Their mathematically optimal decisions often diverge from human expectations, exposing a persistent gap between algorithmic and bounded human rationality. This work addresses this gap by proposing a framework that operationalizes fairness as a learnable objective within optimization-based Digital Twins. We introduce a preference-driven learning pipeline that infers latent fairness objectives directly from human pairwise preferences over feasible decisions. A novel Siamese neural network is developed to generate convex quadratic cost functions conditioned on contextual information. The resulting surrogate objectives align optimization outcomes with human-perceived fairness while maintaining computational efficiency. The approach is demonstrated on a COVID-19 hospital resource allocation scenario. This study provides an actionable path toward embedding human-centered fairness in the design of autonomous decision-making systems.

This paper presents a modality-augmented fine-tuning framework designed to adapt foundation robot policies to diverse humanoid embodiments. We validate our approach across two distinct settings: (i) the GR1 embodiment, utilizing public datasets where we introduce post-processed modalities, including binary contact signals and ZoeDepth-generated metric depth; and (ii) the Unitree G1 embodiment, for which we contribute a novel multi-modal dataset incorporating cuRobo motion planning, inverse kinematics, and ground-truth contact-force measurements. Our experiments demonstrate that modality augmentation consistently enhances policy performance across different embodiments. Specifically, for the GR1, integrating contact-state cues and RGB-D fusion improves online success rates from 51% to 63%. Furthermore, in the G1 "Pick Apple to Bowl" task, our contact-augmented model achieves a success rate of 94%, significantly outperforming the 48% achieved by standard fine-tuning and the 0% baseline of zero-shot transfer. These results highlight that lightweight post-processing effectively strengthens policies for GR1, while high-quality multi-modal data is crucial for reliable transfer to the Unitree G1. Consequently, this work establishes a unified, data-centric pathway for extending foundation robot policies through targeted modality design and multi-modal fine-tuning.

Automated Scientific Discovery (ASD) systems can help automatically generate and run code-based experiments, but their capabilities are limited by the code they can reliably generate from parametric knowledge alone. As a result, current systems either mutate a small number of manually-crafted experiment examples, or operate solely from parametric knowledge, limiting quality and reach. We introduce CodeDistiller, a system that automatically distills large collections of scientific Github repositories into a vetted library of working domain-specific code examples, allowing ASD agents to expand their capabilities without manual effort. Using a combination of automatic and domain-expert evaluation on 250 materials science repositories, we find the best model is capable of producing functional examples for 74% of repositories, while our downstream evaluation shows an ASD agent augmented with a CodeDistiller generated library produces more accurate, complete, and scientifically sound experiments than an agent with only general materials-science code examples.

Agentic large language models are proposed as autonomous code generators for scientific computing, yet their reliability in high-stakes problems remains unclear. Developing computational scientific software from natural-language queries remains challenging broadly due to (a) sparse representation of domain codes during training and (b) the limited feasibility of RLHF with a small expert community. To address these limitations, this work conceptualizes an inverse approach to code design, embodied in the Chain of Unit-Physics framework: a first-principles (or primitives)-centric, multi-agent system in which human expert knowledge is encoded as unit-physics tests that explicitly constrain code generation. The framework is evaluated on a nontrivial combustion task, used here as a representative benchmark for scientific problem with realistic physical constraints. Closed-weight systems and code-focused agentic variants fail to produce correct end-to-end solvers, despite tool and web access, exhibiting four recurrent error classes: interface (syntax/API) hallucinations, overconfident assumptions, numerical/physical incoherence, and configuration fragility. Open-weight models with chain-of-thought (CoT) decoding reduce interface errors but still yield incorrect solutions. On the benchmark task, the proposed framework converges within 5-6 iterations, matches the human-expert implementation (mean error of $3.1\times10^{-3}$ %), with a $\sim$33.4 % faster runtime and a $\sim$30 % efficient memory usage at a cost comparable to mid-sized commercial APIs, yielding a practical template for physics-grounded scientific code generation. As datasets and models evolve, zero-shot code accuracy will improve; however, the Chain of Unit-Physics framework goes further by embedding first-principles analysis that is foundational to scientific codes.

We present Crystalyse, an open, provenance-enforced scientific agent for computational materials design of inorganic crystals that orchestrates tools for compositional screening, crystal structure generation, and machine-learning force-field evaluation. Crystalyse offers three operating modes to trade exploration speed against validation depth: creative (rapid query), adaptive (context-aware routing) and rigorous (comprehensive checks). We release the underlying source code and evaluation scripts to enable plug-and-play use and development. In demonstrations on quaternary oxide exploration, sodium-ion cathode design, and lead-free indoor photovoltaic candidate generation, the agent integrates chemical compound generation with fast stability and property filters. Under adversarial testing, provenance enforcement eliminated material-property hallucinations (a broad adversarial suite pass rate reached 86% from a 57% baseline). Crystalyse provides an agentic artificial intelligence system that can complement existing materials design pipelines, assisting in hypothesis generation while preserving transparency and reproducibility.

Born effective charge (BEC), a fundamental quantity in lattice dynamics and ferroelectric theory, provides a quantitative measure of linear polarization response to ionic displacements. However, it does not account for higher-order effects, which can play a significant role in certain materials, such as fluorite HfO$_2$. In this letter, we extend the BEC framework by introducing the concept of second-order dynamical charge and mode effective charge. Using first-principles calculations, we demonstrate that specific combinations of nonpolar phonon modes in many oxides can induce substantial second-order polarizations, reaching magnitudes comparable to those of intrinsically polar modes. Through a symmetry-based analysis of the charge density, we elucidate the microscopic origin of these effects, tracing them to variations in bond covalency and local electronic rearrangements. We also demonstrate large second-order mode effective charge in well-studied perovskites such as SrTiO$_3$, highlighting the generality of these phenomena. Our results reveal a previously unrecognized mechanism that drives polarization in crystalline solids, offering new insights into the design principles of next-generation ferroelectric, piezoelectric and multifunctional materials.

Graphene has emerged as a promising material for next-generation electronic and thermal devices owing to its exceptional charge transport and thermal conductivity. However, high-quality samples are predominantly obtained via mechanical exfoliation from graphite crystals, a process that inherently lacks scalability. Despite extensive efforts toward large-area synthesis, cost-effective approaches for producing high-quality, large-area, single-crystalline graphene with fast turnaround time remain limited. Here, we report the design, fabrication, and performance benchmarking of a rapid thermal chemical vapor deposition (RTCVD) system capable of synthesizing epitaxial monolayer graphene under atmospheric pressure. The entire growth process, from sample loading to unloading, is achieved within $25$ minutes with a temperature ramp rate exceeding $23^\circ\mathrm{C}/s$. Growth at atmospheric pressure eliminates the need for vacuum components, thereby reducing both system complexity and operational costs. The structural and electronic quality of epitaxial graphene is comprehensively characterized using Raman spectroscopy, selected area electron diffraction (SAED), and magnetotransport measurements, which reveal signatures of quantum Hall effect in synthesized graphene samples. Furthermore, we demonstrate van der Waals epitaxial growth of palladium (Pd) thin films on graphene transferred to Si/SiO$_{2}$ substrates, establishing its single-crystalline nature over a large area and its potential as a versatile platform for subsequent heteroepitaxial growth.

Large language models (LLMs) are increasingly deployed as conversational tutors in STEM education, yet most systems still rely on a single LLM with a static retrieval-augmented generation (RAG) pipeline over course materials. This design struggles in complex domains such as digital signal processing (DSP), where tutors must maintain coherent long-term student models, manage heterogeneous knowledge bases, and adapt teaching strategies over extended interactions. We argue that retrieval, memory, and control should be treated as a coupled cognitive evolution process. We instantiate this view in CogEvo-Edu, a hierarchical educational multi-agent system comprising a Cognitive Perception Layer (CPL), a Knowledge Evolution Layer (KEL), and a Meta-Control Layer (MCL). CPL maintains dual memories and performs confidence-weighted consolidation to build structured, self-correcting student profiles under limited context. KEL assigns each knowledge chunk a spatiotemporal value that drives activation, semantic compression, and forgetting. MCL formulates tutoring as hierarchical sequential decision making, orchestrating specialized agents and jointly adapting CPL/KEL hyperparameters via a dual inner--outer loop. To evaluate CogEvo-Edu, we construct DSP-EduBench, a vertical benchmark for DSP tutoring with heterogeneous resources, simulated student profiles, and long-horizon interaction scripts. Using a three-model LLM-as-a-Judge ensemble, CogEvo-Edu raises the overall score from 5.32 to 9.23 and improves all six indicators over static RAG, simple memory, and a single-agent variant, demonstrating the value of jointly evolving student profiles, knowledge bases, and teaching policies.

Metal-organic frameworks (MOFs) combine high porosity with structural fragility, raising important questions about their mechanical stability. We develop a rigidity-based framework in which spring networks parameterized by UFF4MOF are used to construct rigidity and dynamical matrices. Large-scale analysis of 5,682 MOFs from the CoRE 2019 database shows that most frameworks are formally over-constrained yet cluster sharply near the isostatic threshold, revealing accidental geometric modes and placing many MOFs near mechanical instability. In the representative case of UiO-66, we show that auxiliary long-range constraints introduced by tuning the neighbor cutoff lift these modes into soft, flat, finite-frequency bands. The results show that rigidity-matrix analysis can rapidly identify MOFs likely to remain mechanically stable. This near-criticality mirrors behavior known from topological mechanics and points to a deeper design principle in porous crystals.

The development of next-generation electrochemical energy storage requires devices that combine the high energy density of batteries with the power capability and long cycle life of supercapacitors. However, the interfacial phenomena governing performance in these systems remain poorly unified. The solid-electrolyte interphase (SEI), a nanoscale film formed by electrolyte decomposition, is well studied in batteries but its counterpart in supercapacitors has received limited systematic investigation despite growing experimental evidence. This review argues that SEI formation is a universal electrochemical process that occurs whenever electrode potentials drive electron transfer into electrolyte orbitals beyond their stability limits, independent of whether charge storage is Faradaic or non-Faradaic. Differences between battery SEIs and supercapacitor interphases arise mainly from operating conditions, not fundamental chemistry. Engineered interphases created through electrolyte additives, protective coatings, or surface functionalization suppress leakage currents, improve capacitance retention, and enable stable high-voltage operation. By identifying shared mechanisms and establishing transferable design rules, this unified framework provides a foundation for predictive interphase engineering that supports long-lived, high-performance energy-storage technologies.

Protein inverse folding, the design of an amino acid sequence based on a target 3D structure, is a fundamental problem of computational protein engineering. Existing methods either generate sequences without leveraging external knowledge or relying on protein language models (PLMs). The former omits the evolutionary information stored in protein databases, while the latter is parameter-inefficient and inflexible to adapt to ever-growing protein data. To overcome the above drawbacks, in this paper we propose a novel method, called retrieval-augmented denoising diffusion (RadDiff), for protein inverse folding. Given the target protein backbone, RadDiff uses a hierarchical search strategy to efficiently retrieve structurally similar proteins from large protein databases. The retrieved structures are then aligned residue-by-residue to the target to construct a position-specific amino acid profile, which serves as an evolutionary-informed prior that conditions the denoising process. A lightweight integration module is further designed to incorporate this prior effectively. Experimental results on the CATH, PDB, and TS50 datasets show that RadDiff consistently outperforms existing methods, improving sequence recovery rate by up to 19%. Experimental results also demonstrate that RadDiff generates highly foldable sequences and scales effectively with database size.

The industrial adoption of Artificial Intelligence for Engineering (AI4E) faces two fundamental bottlenecks: scarce high-quality data and the lack of interpretability in black-box models-particularly critical in safety-sensitive sectors like aerospace. We present an explainable, few-shot AI4E framework that is systematically informed by physics and expert knowledge throughout its architecture. Starting from only 32 experimental samples in an aerial K439B superalloy castings repair welding case, we first augment physically plausible synthetic data through a three-stage protocol: differentiated noise injection calibrated to process variabilities, enforcement of hard physical constraints, and preservation of inter-parameter relationships. We then employ a nested optimization strategy for constitutive model discovery, where symbolic regression explores equation structures while differential evolution optimizes parameters, followed by intensive parameter refinement using hybrid global-local optimization. The resulting interpretable constitutive equation achieves 88% accuracy in predicting hot-cracking tendency. This equation not only provides quantitative predictions but also delivers explicit physical insight, revealing how thermal, geometric, and metallurgical mechanisms couple to drive cracking-thereby advancing engineers' cognitive understanding of the process. Furthermore, the constitutive equation serves as a multi-functional tool for process optimization and high-fidelity virtual data generation, enabling accuracy improvements in other data-driven models. Our approach provides a general blueprint for developing trustworthy AI systems that embed engineering domain knowledge directly into their architecture, enabling reliable adoption in high-stakes industrial applications where data is limited but physical understanding is available.

Understanding structure-property relationships in materials is fundamental in condensed matter physics and materials science. Over the past few years, machine learning (ML) has emerged as a powerful tool for advancing this understanding and accelerating materials discovery. Early ML approaches primarily focused on constructing and screening large material spaces to identify promising candidates for various applications. More recently, research efforts have increasingly shifted toward generating crystal structures using end-to-end generative models. This review analyzes the current state of generative modeling for crystal structure prediction and de novo generation. It examines crystal representations, outlines the generative models used to design crystal structures, and evaluates their respective strengths and limitations. Furthermore, the review highlights experimental considerations for evaluating generated structures and provides recommendations for suitable existing software tools. Emerging topics, such as modeling disorder and defects, integration in advanced characterization, incorporating synthetic feasibility constraints, and model explainability are explored. Ultimately, this work aims to inform both experimental scientists looking to adapt suitable ML models to their specific circumstances and ML specialists seeking to understand the unique challenges related to inverse materials design and discovery.

Understanding structure-property relationships in materials is fundamental in condensed matter physics and materials science. Over the past few years, machine learning (ML) has emerged as a powerful tool for advancing this understanding and accelerating materials discovery. Early ML approaches primarily focused on constructing and screening large material spaces to identify promising candidates for various applications. More recently, research efforts have increasingly shifted toward generating crystal structures using end-to-end generative models. This review analyzes the current state of generative modeling for crystal structure prediction and \textit{de novo} generation. It examines crystal representations, outlines the generative models used to design crystal structures, and evaluates their respective strengths and limitations. Furthermore, the review highlights experimental considerations for evaluating generated structures and provides recommendations for suitable existing software tools. Emerging topics, such as modeling disorder and defects, integration in advanced characterization, and incorporating synthetic feasibility constraints, are explored. Ultimately, this work aims to inform both experimental scientists looking to adapt suitable ML models to their specific circumstances and ML specialists seeking to understand the unique challenges related to inverse materials design and discovery.

In living systems, DNA undergoes continuous and rhythmic mechanical remodeling through condensation, looping, and disentangling to regulate gene expression, segregate chromosomes, and guide morphogenesis. Here, we demonstrate a purely mechanical route to rhythmic DNA reorganization in a minimal active composite of microtubules, kinesin motors, and DNA. We embed a DNA polymer in an active turbulent microtubule-kinesin fluid, creating a self-morphing material. The active flows stretch and entangle the DNA, forming a self-organized viscoelastic network that resists active stresses and affects flow over large length scales. This mechanical feedback loop progressively amplifies velocity correlations and drives a nonequilibrium phase transition tuned by DNA contour length: from disordered flow to synchronized, millimeter-scale oscillations with vortices. We rationalize the phase transition with an active-gel model that predicts a growing length scale and an oscillatory instability emerging from the interplay between activity, orientational order, and self-generated viscoelasticity, rather than chemical signaling. The dependence of the oscillation frequency on system size and activity quantitatively agrees with experiment. Thus, flow-driven DNA remodeling provides a minimal physical route to autonomous, system-spanning oscillations in three dimensions and suggests design principles for programmable soft matter that coordinates, actuates, and reshapes itself.

Two-dimensional semiconductors, such as monolayer transition metal dichalcogenides (TMDC), exhibit strong excitonic transitions at room temperature and offer a unique platform for exploring light-matter interactions in nanoscale photonic systems. In this work, we demonstrate a compact and polarization-invariant photonic metasurface, fabricated from hexagonal boron-nitride (hBN) and based on radial bound states in the continuum (BIC), which are formed by radially distributed pairs of structurally asymmetric resonators. The metasurface employs multiple symmetry-breaking perturbations to support high quality-(Q-)factor resonances within a footprint smaller than 8 x 8 $μm^2$ - one-sixth of the area of previous approaches. Compared to established hBN metasurface designs, the radial geometry furthermore achieves significantly higher Q-factors with a reduced footprint. By integrating the hBN photonic structure with a WS$_2$ monolayer, we observe enhanced photoluminescence when its resonance is spectrally aligned with the exciton resonance, accompanied by signatures of discrete momentum-space patterns that identify the orbital-angular-momentum-carrying ring eigenmodes. These features persist over a wide range of excitation powers and show minimal linewidth broadening, indicating robust and spatially modulated exciton-photon coupling. This work establishes a scalable approach for generating hybrid photonic-excitonic states with momentum-space structure, offering new opportunities for exciton localization, valley emission, spatially programmable light-matter interaction in two-dimensional material platforms and compact luminescent devices based on 2D material-integrated metasurfaces.

Designing proteins de novo with tailored structural, physicochemical, and functional properties remains a grand challenge in biotechnology, medicine, and materials science, due to the vastness of sequence space and the complex coupling between sequence, structure, and function. Current state-of-the-art generative methods, such as protein language models (PLMs) and diffusion-based architectures, often require extensive fine-tuning, task-specific data, or model reconfiguration to support objective-directed design, thereby limiting their flexibility and scalability. To overcome these limitations, we present a decentralized, agent-based framework inspired by swarm intelligence for de novo protein design. In this approach, multiple large language model (LLM) agents operate in parallel, each assigned to a specific residue position. These agents iteratively propose context-aware mutations by integrating design objectives, local neighborhood interactions, and memory and feedback from previous iterations. This position-wise, decentralized coordination enables emergent design of diverse, well-defined sequences without reliance on motif scaffolds or multiple sequence alignments, validated with experiments on proteins with alpha helix and coil structures. Through analyses of residue conservation, structure-based metrics, and sequence convergence and embeddings, we demonstrate that the framework exhibits emergent behaviors and effective navigation of the protein fitness landscape. Our method achieves efficient, objective-directed designs within a few GPU-hours and operates entirely without fine-tuning or specialized training, offering a generalizable and adaptable solution for protein design. Beyond proteins, the approach lays the groundwork for collective LLM-driven design across biomolecular systems and other scientific discovery tasks.

Double perovskites (DPs) are promising candidates for sustainable energy technologies due to their compositional tunability and compatibility with low-energy fabrication, yet their vast design space poses a major challenge for conditional materials discovery. This work introduces a multi-agent, text gradient-driven framework that performs DP composition generation under natural-language conditions by integrating three complementary feedback sources: LLM-based self-evaluation, DP-specific domain knowledge-informed feedback, and ML surrogate-based feedback. Analogous to how knowledge-informed machine learning improves the reliability of conventional data-driven models, our framework incorporates domain-informed text gradients to guide the generative process toward physically meaningful regions of the DP composition space. Systematic comparison of three incremental configurations, (i) pure LLM generation, (ii) LLM generation with LLM reasoning-based feedback, and (iii) LLM generation with domain knowledge-guided feedback, shows that iterative guidance from knowledge-informed gradients improves stability-condition satisfaction without additional training data, achieving over 98% compositional validity and up to 54% stable or metastable candidates, surpassing both the LLM-only baseline (43%) and prior GAN-based results (27%). Analyses of ML-based gradients further reveal that they enhance performance in in-distribution (ID) regions but become unreliable in out-of-distribution (OOD) regimes. Overall, this work provides the first systematic analysis of multi-agent, knowledge-guided text gradients for DP discovery and establishes a generalizable blueprint for MAS-driven generative materials design aimed at advancing sustainable technologies.

Double perovskites (DPs) are promising candidates for sustainable energy technologies due to their compositional tunability and compatibility with low-energy fabrication, yet their vast design space poses a major challenge for conditional materials discovery. This work introduces a multi-agent, text gradient-driven framework that performs DP composition generation under natural-language conditions by integrating three complementary feedback sources: LLM-based self-evaluation, DP-specific domain knowledge-informed feedback, and ML surrogate-based feedback. Analogous to how knowledge-informed machine learning improves the reliability of conventional data-driven models, our framework incorporates domain-informed text gradients to guide the generative process toward physically meaningful regions of the DP composition space. Systematic comparison of three incremental configurations, (i) pure LLM generation, (ii) LLM generation with LLM reasoning-based feedback, and (iii) LLM generation with domain knowledge-guided feedback, shows that iterative guidance from knowledge-informed gradients improves stability-condition satisfaction without additional training data, achieving over 98% compositional validity and up to 54% stable or metastable candidates, surpassing both the LLM-only baseline (43%) and prior GAN-based results (27%). Analyses of ML-based gradients further reveal that they enhance performance in in-distribution (ID) regions but become unreliable in out-of-distribution (OOD) regimes. Overall, this work provides the first systematic analysis of multi-agent, knowledge-guided text gradients for DP discovery and establishes a generalizable blueprint for MAS-driven generative materials design aimed at advancing sustainable technologies.

Experimental validation of chemical processes is slow and costly, limiting exploration in materials discovery. Machine learning can prioritize promising candidates, but existing data in patents and literature is heterogeneous and difficult to use. We introduce a universal directed-tree process-graph representation that unifies unstructured text, molecular structures, and numeric measurements into a single machine-readable format. To learn from this structured data, we developed a multi-modal graph neural network with a property-conditioned attention mechanism. Trained on approximately 700,000 process graphs from nearly 9,000 diverse documents, our model learns semantically rich embeddings that generalize across domains. When fine-tuned on compact, domain-specific datasets, the pretrained model achieves strong performance, demonstrating that universal process representations learned at scale transfer effectively to specialized prediction tasks with minimal additional data.

Generative models have recently shown great promise for accelerating the design and discovery of new functional materials. Conditional generation enhances this capacity by allowing inverse design, where specific desired properties can be requested during the generation process. However, conditioning of transformer-based approaches, in particular, is constrained by discrete tokenisation schemes and the risk of catastrophic forgetting during fine-tuning. This work introduces CrystaLLM-π (property injection), a conditional autoregressive framework that integrates continuous property representations directly into the transformer's attention mechanism. Two architectures, Property-Key-Value (PKV) Prefix attention and PKV Residual attention, are presented. These methods bypass inefficient sequence-level tokenisation and preserve foundational knowledge from unsupervised pre-training on Crystallographic Information Files (CIFs) as textual input. We establish the efficacy of these mechanisms through systematic robustness studies and evaluate the framework's versatility across two distinct tasks. First, for structure recovery, the model processes high-dimensional, heterogeneous X-ray diffraction patterns, achieving structural accuracy competitive with specialised models and demonstrating applications to experimental structure recovery and polymorph differentiation. Second, for materials discovery, the model is fine-tuned on a specialised photovoltaic dataset to generate novel, stable candidates validated by Density Functional Theory (DFT). It implicitly learns to target optimal band gap regions for high photovoltaic efficiency, demonstrating a capability to map complex structure-property relationships. CrystaLLM-π provides a unified, flexible, and computationally efficient framework for inverse materials design.

High-Entropy/multicomponent rare-earth oxides (HECs and MCCs) show promise as alternative materials for thermal barrier coatings (TBC) with the ability to tailor properties based on the combination of rare-earth elements present. By enabling the substitution of scarce or supply-risk rare-earths with more readily available alternatives while maintaining comparable material performance, HECs and MCCs offer a valuable path towards alternative TBC material design. However, navigating this search space of compositionally complex materials is both time and resource intensive. In this study, an active learning (AL) framework was employed to identify HEC/MCC materials with a pyrochlore structure, with acceptable thermal conductivity (TC) for TBC applications. The AL framework was applied through a Bayesian optimisation (BO) strategy, coupled with a random forest surrogate model. TC was selected as the optimisation criterion as that is the most basic requirement of TBC materials. Over two iterations of the AL cycle, four compositions were generated and synthesized in the lab for experimental evaluation. The first iteration yielded two single-phase pyrochlores, $(La_{0.29}Nd_{0.36}Gd_{0.36})_2Zr_2O_7$ and $(La_{0.333}Nd_{0.26}Gd_{0.15}Ho_{0.15}Yb_{0.111})_2Zr_2O_7$, with measured thermal conductivities of 2.03 and 1.90 $W/mK$, respectively. The surrogate model predicted a TC of 2.009 $W/mK$ for both compositions, demonstrating it's accuracy for completely new compositions. The second iteration compositions showed dual-phase when synthesized, highlighting the need to take into account phase formation in the AL framework.

Artificial intelligence and machine learning are reshaping how we approach scientific discovery, not by replacing established methods but by extending what researchers can probe, predict, and design. In this roadmap we provide a forward-looking view of AI-enabled science across biology, chemistry, climate science, mathematics, materials science, physics, self-driving laboratories and unconventional computing. Several shared themes emerge: the need for diverse and trustworthy data, transferable electronic-structure and interatomic models, AI systems integrated into end-to-end scientific workflows that connect simulations to experiments and generative systems grounded in synthesisability rather than purely idealised phases. Across domains, we highlight how large foundation models, active learning and self-driving laboratories can close loops between prediction and validation while maintaining reproducibility and physical interpretability. Taken together, these perspectives outline where AI-enabled science stands today, identify bottlenecks in data, methods and infrastructure, and chart concrete directions for building AI systems that are not only more powerful but also more transparent and capable of accelerating discovery in complex real-world environments.

The advancement of solid-state batteries depends on the development of lithium-ion conductors that exhibit both high ionic conductivity and stability across a wide range of electrochemical and chemical conditions. In this paper, we investigate the chemical factors that control the stability of Li-NASICONs and garnets in highly alkaline aqueous environment. While this is of general importance, it is particularly important for the operation of Li-air cells with humidified air. Humid air promotes the formation of LiOH as the discharge product, creating a highly alkaline environment on the surface of cathode and solid-state electrolyte. In this work, we combine machine learning and first-principles calculations to conduct a high-throughput computational screening of alkaline-stable oxide-based Li-ion conductors in order to better characterize the tradeoff between the various relevant properties. We evaluate the material stability in terms of pH, voltage, and species present in the environment (LiOH and H2O) across a vast range of chemical compositions with NASICON and garnet crystal structures. We utilize the CHGNet universal machine learning interatomic potential for pre-screening, followed by DFT calculations. Such a hierarchical screening procedure enables the evaluation of over 320,000 chemical compositions, encompassing nearly the entire periodic table. From this set 209 alkaline-stable NASICON and garnet compounds are selected as final candidates. We identify the specific cation substitutions that improve alkaline stability in NASICON and garnet compounds, and reveal the underlying mechanism. We also discover the trade-offs for designing alkaline-stable Li-ion conductors, highlighting the need to carefully optimize compositions so that it can simultaneously enhance all the material properties required for practical battery applications.

Mechanical design and manufacturing workflows conventionally begin with conceptual design, followed by the creation of a computer-aided design (CAD) model and fabrication through material-extrusion (MEX) printing. This process requires converting CAD geometry into machine-readable G-code through slicing and path planning. While each step is well established, dependence on CAD modeling remains a major bottleneck: constructing object-specific 3D geometry is slow and poorly suited to rapid prototyping. Even minor design variations typically necessitate manual updates in CAD software, making iteration time-consuming and difficult to scale. To address this limitation, we introduce Image2Gcode, an end-to-end data-driven framework that bypasses the CAD stage and generates printer-ready G-code directly from images and part drawings. Instead of relying on an explicit 3D model, a hand-drawn or captured 2D image serves as the sole input. The framework first extracts slice-wise structural cues from the image and then employs a denoising diffusion probabilistic model (DDPM) over G-code sequences. Through iterative denoising, the model transforms Gaussian noise into executable print-move trajectories with corresponding extrusion parameters, establishing a direct mapping from visual input to native toolpaths. By producing structured G-code directly from 2D imagery, Image2Gcode eliminates the need for CAD or STL intermediates, lowering the entry barrier for additive manufacturing and accelerating the design-to-fabrication cycle. This approach supports on-demand prototyping from simple sketches or visual references and integrates with upstream 2D-to-3D reconstruction modules to enable an automated pipeline from concept to physical artifact. The result is a flexible, computationally efficient framework that advances accessibility in design iteration, repair workflows, and distributed manufacturing.

Reasoning based on Large Language Models (LLMs) has garnered increasing attention due to outstanding performance of these models in mathematical and complex logical tasks. Beginning with the Chain-of-Thought (CoT) prompting technique, numerous reasoning methods have emerged that decompose problems into smaller, sequential steps (or thoughts). However, existing reasoning models focus on conventional problem-solving and do not necessarily generate creative solutions by ``creative reasoning''. In domains where the solution space is expansive and conventional solutions are suboptimal, such as drug discovery or business strategization, creative reasoning to discover innovative solutions is crucial. To address this gap, first we introduce a computational framework for creative reasoning inspired by established cognitive science principles. With this framework, we propose three core creative reasoning paradigms, namely, \textit{combinational}, \textit{exploratory}, and \textit{transformative} reasoning, where each offers specific directions for systematic exploration of the universe of thoughts to generate creative solutions. Next, to materialize this framework using LLMs, we introduce the \textit{Universe of Thoughts} (or \textit{UoT}, for short), a novel set of methods to implement the aforementioned three creative processes. Finally, we introduce three novel tasks that necessitate creative problem-solving, along with an evaluation benchmark to assess creativity from three orthogonal perspectives: feasibility as constraint, and utility and novelty as metrics. With a comparative analysis against the state-of-the-art (SOTA) reasoning techniques as well as representative commercial models with reasoning capability, we show that UoT demonstrates superior performance in creative reasoning.

Reasoning based on Large Language Models (LLMs) has garnered increasing attention due to outstanding performance of these models in mathematical and complex logical tasks. Beginning with the Chain-of-Thought (CoT) prompting technique, numerous reasoning methods have emerged that decompose problems into smaller, sequential steps (or thoughts). However, existing reasoning models focus on conventional problem-solving and do not necessarily generate creative solutions by ``creative reasoning''. In domains where the solution space is expansive and conventional solutions are suboptimal, such as drug discovery or business strategization, creative reasoning to discover innovative solutions is crucial. To address this gap, first we introduce a computational framework for creative reasoning inspired by established cognitive science principles. With this framework, we propose three core creative reasoning paradigms, namely, \textit{combinational}, \textit{exploratory}, and \textit{transformative} reasoning, where each offers specific directions for systematic exploration of the universe of thoughts to generate creative solutions. Next, to materialize this framework using LLMs, we introduce the \textit{Universe of Thoughts} (or \textit{UoT}, for short), a novel set of methods to implement the aforementioned three creative processes. Finally, we introduce three novel tasks that necessitate creative problem-solving, along with an evaluation benchmark to assess creativity from three orthogonal perspectives: feasibility as constraint, and utility and novelty as metrics. With a comparative analysis against the state-of-the-art (SOTA) reasoning techniques as well as representative commercial models with reasoning capability, we show that UoT demonstrates superior performance in creative reasoning.

Designing materials with controlled heat flow at the nano-scale is central to advances in microelectronics, thermoelectrics, and energy-conversion technologies. At these scales, phonon transport follows the Boltzmann Transport Equation (BTE), which captures non-diffusive (ballistic) effects but is too costly to solve repeatedly in inverse-design loops. Existing surrogate approaches trade speed for accuracy: fast macroscopic solvers can overestimate conductivities by hundreds of percent, while recent data-driven operator learners often require thousands of high-fidelity simulations. This creates a need for a fast, data-efficient surrogate that remains reliable across ballistic and diffusive regimes. We introduce a Physics-Enhanced Deep Surrogate (PEDS) that combines a differentiable Fourier solver with a neural generator and couples it with uncertainty-driven active learning. The Fourier solver acts as a physical inductive bias, while the network learns geometry-dependent corrections and a mixing coefficient that interpolates between macroscopic and nano-scale behavior. PEDS reduces training-data requirements by up to 70% compared with purely data-driven baselines, achieves roughly 5% fractional error with only 300 high-fidelity BTE simulations, and enables efficient design of porous geometries spanning 12-85 W m$^{-1}$ K$^{-1}$ with average design errors of 4%. The learned mixing parameter recovers the ballistic-diffusive transition and improves out of distribution robustness. These results show that embedding simple, differentiable low-fidelity physics can dramatically increase surrogate data-efficiency and interpretability, making repeated PDE-constrained optimization practical for nano-scale thermal-materials design.

Stellarators are a prospective class of fusion-based power plants that confine a hot plasma with three-dimensional magnetic fields. Typically framed as a PDE-constrained optimization problem, stellarator design is a time-consuming process that can take hours to solve on a computing cluster. Developing fast methods for designing stellarators is crucial for advancing fusion research. Given the recent development of large datasets of optimized stellarators, machine learning approaches have emerged as a potential candidate. Motivated by this, we present an open inverse problem to the machine learning community: to rapidly generate high-quality stellarator designs which have a set of desirable characteristics. As a case study in the problem space, we train a conditional diffusion model on data from the QUASR database to generate quasisymmetric stellarator designs with desirable characteristics (aspect ratio and mean rotational transform). The diffusion model is applied to design stellarators with characteristics not seen during training. We provide evaluation protocols and show that many of the generated stellarators exhibit solid performance: less than 5% deviation from quasisymmetry and the target characteristics. The modest deviation from quasisymmetry highlights an opportunity to reach the sub 1% target. Beyond the case study, we share multiple promising avenues for generative modeling to advance stellarator design.

Radio maps (RMs) serve as environment-aware electromagnetic (EM) representations that connect scenario geometry and material properties to the spatial distribution of signal strength, enabling localization without costly in-situ measurements. However, constructing high-fidelity indoor RMs remains challenging due to the prohibitive latency of EM solvers and the limitations of learning-based methods, which often rely on sparse measurements or assumptions of homogeneous material, which are misaligned with the heterogeneous and multipath-rich nature of indoor environments. To overcome these challenges, we propose iRadioDiff, a sampling-free diffusion-based framework for indoor RM construction. iRadioDiff is conditioned on access point (AP) positions, and physics-informed prompt encoded by material reflection and transmission coefficients. It further incorporates multipath-critical priors, including diffraction points, strong transmission boundaries, and line-of-sight (LoS) contours, to guide the generative process via conditional channels and boundary-weighted objectives. This design enables accurate modeling of nonstationary field discontinuities and efficient construction of physically consistent RMs. Experiments demonstrate that iRadioDiff achieves state-of-the-art performance in indoor RM construction and received signal strength based indoor localization, which offers effective generalization across layouts and material configurations. Code is available at https://github.com/UNIC-Lab/iRadioDiff.

Vacancy engineering, the intentional control of atomic-scale vacancies in metals and alloys, is emerging as a powerful yet underexplored strategy for tailoring microstructures and optimizing performance across diverse applications. By enabling excess vacancy populations through quenching, severe deformation, thermomechanical treatments, or additive manufacturing, new microstructures can be obtained that achieve unique combinations of strength, ductility, fatigue life, corrosion resistance, and conductivity. Vacancies are distinct among lattice defects: they are non-conserved entities essential for solute diffusion, yet variably coupled to solutes, dislocations, and phase boundaries. They can accelerate transformations such as nucleation and precipitation or retard kinetics when trapped in clusters, and their transient trapping and release can drive microstructural evolution across time and length scales. This Review synthesizes recent advances in generating, modeling, and characterizing vacancies, highlighting their role in diffusion, precipitation, and phase stability. Case studies in lightweight, high-temperature, fatigue-resistant, electrical, and biomedical materials demonstrate the broad potential of vacancy control. We conclude by emphasizing the opportunity for the metallurgical community to fully exploit excess vacancies as controllable, design-relevant defects that enable new pathways for microstructure and property optimization in next-generation alloys.

Vacancy engineering, the intentional control of atomic-scale vacancies in metals and alloys, is emerging as a powerful yet underexplored strategy for tailoring microstructures and optimizing performance across diverse applications. By enabling excess vacancy populations through quenching, severe deformation, thermomechanical treatments, or additive manufacturing, new microstructures can be obtained that achieve unique combinations of strength, ductility, fatigue life, corrosion resistance, and conductivity. Vacancies are distinct among lattice defects: they are non-conserved entities essential for solute diffusion, yet variably coupled to solutes, dislocations, and phase boundaries. They can accelerate transformations such as nucleation and precipitation or retard kinetics when trapped in clusters, and their transient trapping and release can drive microstructural evolution across time and length scales. This Review synthesizes recent advances in generating, modeling, and characterizing vacancies, highlighting their role in diffusion, precipitation, and phase stability. Case studies in lightweight, high-temperature, fatigue-resistant, electrical, and biomedical materials demonstrate the broad potential of vacancy control. We conclude by emphasizing the opportunity for the metallurgical community to fully exploit excess vacancies as controllable, design-relevant defects that enable new pathways for microstructure and property optimization in next-generation alloys.

The discovery of high-performance photosensitizers has long been hindered by the time-consuming and resource-intensive nature of traditional trial-and-error approaches. Here, we present \textbf{A}I-\textbf{A}ccelerated \textbf{P}hoto\textbf{S}ensitizer \textbf{I}nnovation (AAPSI), a closed-loop workflow that integrates expert knowledge, scaffold-based molecule generation, and Bayesian optimization to accelerate the design of novel photosensitizers. The scaffold-driven generation in AAPSI ensures structural novelty and synthetic feasibility, while the iterative AI-experiment loop accelerates the discovery of novel photosensitizers. AAPSI leverages a curated database of 102,534 photosensitizer-solvent pairs and generate 6,148 synthetically accessible candidates. These candidates are screened via graph transformers trained to predict singlet oxygen quantum yield ($φ_Δ$) and absorption maxima ($λ_{max}$), following experimental validation. This work generates several novel candidates for photodynamic therapy (PDT), among which the hypocrellin-based candidate HB4Ph exhibits exceptional performance at the Pareto frontier of high quantum yield of singlet oxygen and long absorption maxima among current photosensitizers ($φ_Δ$=0.85, $λ_{max}$=650nm).

High-throughput methods enable accelerated discovery of novel materials in complex systems such as high-entropy alloys, which exhibit intricate phase stability across vast compositional spaces. Computational approaches, including Density Functional Theory (DFT) and calculation of phase diagrams (CALPHAD), facilitate screening of phase formability as a function of composition and temperature. However, the integration of computational predictions with experimental validation remains challenging in high-throughput studies. In this work, we introduce a quantitative confidence metric to assess the agreement between predictions and experimental observations, providing a quantitative measure of the confidence of machine learning models trained on either DFT or CALPHAD input in accounting for experimental evidence. The experimental dataset was generated via high-throughput in-situ synchrotron X-ray diffraction on compositionally varied FeNiMnCr alloy libraries, heated from room temperature to ~1000 °C. Agreement between the observed and predicted phases was evaluated using either temperature-independent phase classification or a model that incorporates a temperature-dependent probability of phase formation. This integrated approach demonstrates where strong overall agreement between computation and experiment exists, while also identifying key discrepancies, particularly in FCC/BCC predictions at Mn-rich regions to inform future model refinement.

Generative Flow Networks, or GFlowNets, offer a promising framework for molecular design, but their internal decision policies remain opaque. This limits adoption in drug discovery, where chemists require clear and interpretable rationales for proposed structures. We present an interpretability framework for SynFlowNet, a GFlowNet trained on documented chemical reactions and purchasable starting materials that generates both molecules and the synthetic routes that produce them. Our approach integrates three complementary components. Gradient based saliency combined with counterfactual perturbations identifies which atomic environments influence reward and how structural edits change molecular outcomes. Sparse autoencoders reveal axis aligned latent factors that correspond to physicochemical properties such as polarity, lipophilicity, and molecular size. Motif probes show that functional groups including aromatic rings and halogens are explicitly encoded and linearly decodable from the internal embeddings. Together, these results expose the chemical logic inside SynFlowNet and provide actionable and mechanistic insight that supports transparent and controllable molecular design.

Perovskite solar cells are promising candidates for next-generation photovoltaics. However, their performance as multi-scale devices is determined by complex interactions between their constituent layers. This creates a vast combinatorial space of possible materials and device architectures, making the conventional experimental-based screening process slow and expensive. Machine learning models try to address this problem, but they only focus on individual material properties or neglect the important geometric information of the perovskite crystal. To address this problem, we propose to predict perovskite solar cell power conversion efficiency with a geometric-aware co-attention (Solar-GECO) model. Solar-GECO combines a geometric graph neural network (GNN) - that directly encodes the atomic structure of the perovskite absorber - with language model embeddings that process the textual strings representing the chemical compounds of the transport layers and other device components. Solar-GECO also integrates a co-attention module to capture intra-layer dependencies and inter-layer interactions, while a probabilistic regression head predicts both power conversion efficiency (PCE) and its associated uncertainty. Solar-GECO achieves state-of-the-art performance, significantly outperforming several baselines, reducing the mean absolute error (MAE) for PCE prediction from 3.066 to 2.936 compared to semantic GNN (the previous state-of-the-art model). Solar-GECO demonstrates that integrating geometric and textual information provides a more powerful and accurate framework for PCE prediction.

In finance, predictive models must balance accuracy and interpretability, particularly in credit risk assessment, where model decisions carry material consequences. We present a quantum neural network (QNN) based on a single qudit, in which both data features and trainable parameters are co-encoded within a unified unitary evolution generated by the full Lie algebra. This design explores the entire Hilbert space while enabling interpretability through the magnitudes of the learned coefficients. We benchmark our model on a real-world, imbalanced credit-risk dataset from Taiwan. The proposed QNN consistently outperforms LR and reaches the results of random forest models in macro-F1 score while preserving a transparent correspondence between learned parameters and input feature importance. To quantify the interpretability of the proposed model, we introduce two complementary metrics: (i) the edit distance between the model's feature ranking and that of LR, and (ii) a feature-poisoning test where selected features are replaced with noise. Results indicate that the proposed quantum model achieves competitive performance while offering a tractable path toward interpretable quantum learning.

Foundation models (FM) are transforming artificial intelligence by enabling generalizable, data-efficient solutions across different domains for a broad range of applications. However, the lack of large and diverse datasets limits the development of FM in nanophotonics. This work presents MOCLIP (Metasurface Optics Contrastive Learning Pretrained), a nanophotonic foundation model that integrates metasurface geometry and spectra within a shared latent space. MOCLIP employs contrastive learning to align geometry and spectral representations using an experimentally acquired dataset with a sample density comparable to ImageNet-1K. The study demonstrates MOCLIP inverse design capabilities for high-throughput zero-shot prediction at a rate of 0.2 million samples per second, enabling the design of a full 4-inch wafer populated with high-density metasurfaces in minutes. It also shows generative latent-space optimization reaching 97 percent accuracy. Finally, we introduce an optical information storage concept that uses MOCLIP to achieve a density of 0.1 Gbit per square millimeter at the resolution limit, exceeding commercial optical media by a factor of six. These results position MOCLIP as a scalable and versatile platform for next-generation photonic design and data-driven applications.

In crystalline solids, the electronic polarization follows the \emph{generalized Neumann's principle}, under which all crystallographic point groups can, in principle, support ferroelectric polarization. However, in high-symmetry structures, polarization is constrained by symmetry operations and becomes quantized into discrete values. We demonstrate that this quantized polarization (QP) is not a mathematical artifact but a \emph{symmetry-protected invariant} that encodes intrinsic information about a material's symmetry and electronic structure. Because of its discrete and non-continuous nature, when two materials with different QPs form an interface, their bulk polarization states cannot be connected adiabatically, compelling the system to develop pronounced interfacial responses: such as metallic states, bound charges, or strong lattice distortions. This theoretical framework provides a unified reinterpretation of classical systems such as the LaAlO$_3$/SrTiO$_3$ interface, revealing it as a prototypical case of QP mismatch. By establishing QP as a fundamental bulk invariant, our work uncovers a universal mechanism governing interfacial electronic phenomena and opens new pathways for the design of functional quantum materials through engineered polarization mismatch.

Electric polarization in the absence of an externally applied electric field is a key property of polar materials, but the standard interpolation-based ab initio approach to compute polarization differences within the modern theory of polarization presents challenges for automated high-throughput calculations. Berry flux diagonalization [J. Bonini et. al, Phys. Rev. B 102, 045141 (2020)] has been proposed as an efficient and reliable alternative, though it has yet to be widely deployed. Here, we assess Berry flux diagonalization using ab initio calculations of a large set of materials, introducing and validating heuristics that ensure branch alignment with a minimal number of intermediate interpolated structures. Our automated implementation of Berry flux diagonalization succeeds in cases where prior interpolation-based workflows fail due to band-gap closures or branch ambiguities. Benchmarking with ab initio calculations of 176 candidate ferroelectrics, we demonstrate the efficacy of the approach on a broad range of insulating materials and obtain accurate effective polarization values with fewer interpolated structures than prior automated interpolation-based workflows. Our real-space heuristics that can predict gauge stability a priori from ionic displacements enable a general automated framework for reliable polarization calculations and efficient high-throughput screening of chemically and structurally diverse polar insulators. These results establish Berry flux diagonalization as a robust and efficient method to compute the effective polarization of solids and to accelerate the data-driven discovery of functional polar materials.

Carbon nanotube (CNT)-polymer composites are widely engineered using surface coatings and chemical treatments to improve interfacial bonding and load transfer. It has been suggested in the nanocomposite literature that nanotube curvature, in conjunction with surface functionalization such as polydopamine (PDA) coating, could serve as an additional control knob for tuning interfacial bonding and energy dissipation in polymer-CNT systems. While experimental and simulation studies have demonstrated the benefits of PDA functionalization, the fundamental mechanism by which nanotube curvature modulates interfacial energy flow and mechanical polarity remains unresolved. This gap is sharpened by a persistent paradox: identical PDA functionalization strengthens some CNT-polymer systems while weakening others, a curvature-dependent inconsistency that has remained unexplained. Here, we employ reactive molecular dynamics (ReaxFF) simulations to resolve how curvature and PDA functionalization jointly govern interfacial energy evolution in CNT-polyvinyl alcohol (PVA) nanocomposites. Our investigation reveals that curvature and PDA functionalization jointly produce opposite regimes of interfacial energy flow: high-curvature CNTs generate dissipative, frictional interphases, whereas low-curvature CNTs confine energy in rigid, cohesive shells. This polarity inversion originates from a curvature-induced transition in PDA adsorption geometry that transforms the interphase from an energy-releasing to an energy-storing configuration. These results establish curvature as a fundamental design parameter for engineering polymer-nanotube interfaces, offering a predictive route to tune interfacial energy flow, mechanical resilience, and transport properties beyond the limits of conventional chemical functionalization.

Organic photovoltaic (OPV) materials offer a promising path toward sustainable energy generation, but their development is limited by the difficulty of identifying high performance donor and acceptor pairs with strong power conversion efficiencies (PCEs). Existing design strategies typically focus on either the donor or the acceptor alone, rather than using a unified approach capable of modeling both components. In this work, we introduce a dual machine learning framework for OPV discovery that combines predictive modeling with generative molecular design. We present the Organic Photovoltaic Donor Acceptor Dataset (OPV2D), the largest curated dataset of its kind, containing 2000 experimentally characterized donor acceptor pairs. Using this dataset, we develop the Organic Photovoltaic Classifier (OPVC) to predict whether a material exhibits OPV behavior, and a hierarchical graph neural network that incorporates multi task learning and donor acceptor interaction modeling. This framework includes the Molecular Orbital Energy Estimator (MOE2) for predicting HOMO and LUMO energy levels, and the Photovoltaic Performance Predictor (P3) for estimating PCE. In addition, we introduce the Material Generative Pretrained Transformer (MatGPT) to produce synthetically accessible organic semiconductors, guided by a reinforcement learning strategy with three objective policy optimization. By linking molecular representation learning with performance prediction, our framework advances data driven discovery of high performance OPV materials.

We present a long-horizon, hierarchical deep research (DR) agent designed for complex materials and device discovery problems that exceed the scope of existing Machine Learning (ML) surrogates and closed-source commercial agents. Our framework instantiates a locally deployable DR instance that integrates local retrieval-augmented generation with large language model reasoners, enhanced by a Deep Tree of Research (DToR) mechanism that adaptively expands and prunes research branches to maximize coverage, depth, and coherence. We systematically evaluate across 27 nanomaterials/device topics using a large language model (LLM)-as-judge rubric with five web-enabled state-of-the-art models as jurors. In addition, we conduct dry-lab validations on five representative tasks, where human experts use domain simulations (e.g., density functional theory, DFT) to verify whether DR-agent proposals are actionable. Results show that our DR agent produces reports with quality comparable to--and often exceeding--those of commercial systems (ChatGPT-5-thinking/o3/o4-mini-high Deep Research) at a substantially lower cost, while enabling on-prem integration with local data and tools.

We present a long-horizon, hierarchical deep research (DR) agent designed for complex materials and device discovery problems that exceed the scope of existing Machine Learning (ML) surrogates and closed-source commercial agents. Our framework instantiates a locally deployable DR instance that integrates local retrieval-augmented generation with large language model reasoners, enhanced by a Deep Tree of Research (DToR) mechanism that adaptively expands and prunes research branches to maximize coverage, depth, and coherence. We systematically evaluate across 27 nanomaterials/device topics using a large language model (LLM)-as-judge rubric with five web-enabled state-of-the-art models as jurors. In addition, we conduct dry-lab validations on five representative tasks, where human experts use domain simulations (e.g., density functional theory, DFT) to verify whether DR-agent proposals are actionable. Results show that our DR agent produces reports with quality comparable to--and often exceeding--those of commercial systems (ChatGPT-5-thinking/o3/o4-mini-high Deep Research) at a substantially lower cost, while enabling on-prem integration with local data and tools.

The discovery of next-generation photoinitiators for two-photon polymerization (TPP) is hindered by the absence of large, open datasets containing the quantum-chemical and photophysical properties required to model photodissociation and excited-state behavior. Existing molecular datasets typically provide only basic physicochemical descriptors and therefore cannot support data-driven screening or AI-assisted design of photoinitiators. To address this gap, we introduce QuantumChem-200K, a large-scale dataset of over 200,000 organic molecules annotated with eleven quantum-chemical properties, including two-photon absorption (TPA) cross sections, TPA spectral ranges, singlet-triplet intersystem crossing (ISC) energies, toxicity and synthetic accessibility scores, hydrophilicity, solubility, boiling point, molecular weight, and aromaticity. These values are computed using a hybrid workflow that integrates density function theory (DFT), semi-empirical excited-state methods, atomistic quantum solvers, and neural-network predictors. Using QuantumChem-200K, we fine tune the open-source Qwen2.5-32B large language model to create a chemistry AI assistant capable of forward property prediction from SMILES. Benchmarking on 3000 unseen molecules from VQM24 and ZINC20 demonstrates that domain-specific fine-tuning significantly improves accuracy over GPT-4o, Llama-3.1-70B, and the base Qwen2.5-32B model, particularly for TPA and ISC predictions central to photoinitiator design. QuantumChem-200K and the corresponding AI assistant together provide the first scalable platform for high-throughput, LLM-driven photoinitiator screening and accelerated discovery of photosensitive materials.

In materials discovery, the integration of first-principles calculations with machine learning techniques has been actively studied for two key tasks: crystal structure prediction, which searches for stable structures given a chemical composition, and elemental substitution, which explores chemical compositions that yield desirable properties in a given crystal structure. However, even when both the crystal structure and chemical composition are fixed, material properties can still vary depending on the atomic arrangements (configurations) at crystallographic sites. To support detailed material design, we present PyAPX, a Python toolkit that performs Bayesian searches of stable atomic configurations. A distinctive feature of this initial release is the introduction of encoding methods suitable for configuration search, and we evaluate their performance using the h-BCN system. As a result, they were confirmed to yield superior convergence compared to commonly used one-hot encoding. PyAPX is broadly applicable to crystalline materials and is expected to further advance materials discovery.

Natural hyperbolic materials have attracted significant interest in the field of photonics due to their unique optical properties. Based on the initial successful explorations on layered crystalline materials, hyperbolic dispersion was associated with extreme structural anisotropy, despite the rarity of natural materials exhibiting this property. Here we show that non cubic electrides are generally promising natural hyperbolic materials owing to charge localization in interstitial sites. This includes elemental and binary electrides, as well as some two-dimensional materials that show prominent in-plane hyperbolic dispersion. They exhibit low plasma frequencies and a broad hyperbolic window spanning the infrared to the ultraviolet. In semiconductor electrides, anisotropic interband transitions provide an additional mechanism for hyperbolic behaviour. These findings remove the previously held prerequisite of structural anisotropy for natural hyperbolic materials, and open up new opportunities, which might change the current strategy for searching and design photonic materials.

Machine learning (ML) has emerged as a powerful tool for accelerating the computational design and production of materials. In materials science, ML has primarily supported large-scale discovery of novel compounds using first-principles data and digital twin applications for optimizing manufacturing processes. However, applying general-purpose ML frameworks to complex industrial materials such as steel remains a challenge. A key obstacle is accurately capturing the intricate relationship between chemical composition, processing parameters, and the resulting microstructure and properties. To address this, we introduce a computational framework that combines physical insights with ML to develop a physics-informed continuous cooling transformation (CCT) model for steels. Our model, trained on a dataset of 4,100 diagrams, is validated against literature and experimental data. It demonstrates high computational efficiency, generating complete CCT diagrams with 100 cooling curves in under 5 seconds. It also shows strong generalizability across alloy steels, achieving phase classification F1 scores above 88% for all phases. For phase transition temperature regression, it attains mean absolute errors (MAE) below 20 °C across all phases except bainite, which shows a slightly higher MAE of 27 °C. This framework can be extended with additional generic and customized ML models to establish a universal digital twin platform for heat treatment. Integration with complementary simulation tools and targeted experiments will further support accelerated materials design workflows.

Efficiently and meaningfully estimating prediction uncertainty is important for exploration in active learning campaigns in materials discovery, where samples with high uncertainty are interpreted as containing information missing from the model. In this work, the effect of different uncertainty estimation and calibration methods are evaluated for active learning when using ensembles of ALIGNN, eXtreme Gradient Boost, Random Forest, and Neural Network model architectures. We compare uncertainty estimates from ALIGNN deep ensembles to loss landscape uncertainty estimates obtained for solubility, bandgap, and formation energy prediction tasks. We then evaluate how the quality of the uncertainty estimate impacts an active learning campaign that seeks model generalization to out-of-distribution data. Uncertainty calibration methods were found to variably generalize from in-domain data to out-of-domain data. Furthermore, calibrated uncertainties were generally unsuccessful in reducing the amount of data required by a model to improve during an active learning campaign on out-of-distribution data when compared to random sampling and uncalibrated uncertainties. The impact of poor-quality uncertainty persists for random forest and eXtreme Gradient Boosting models trained on the same data for the same tasks, indicating that this is at least partially intrinsic to the data and not due to model capacity alone. Analysis of the target, in-distribution uncertainty, out-of-distribution uncertainty, and training residual distributions suggest that future work focus on understanding empirical uncertainties in the feature input space for cases where ensemble prediction variances do not accurately capture the missing information required for the model to generalize.

High-resolution (HR) magnetic resonance imaging (MRI) is crucial for many clinical and research applications. However, achieving it remains costly and constrained by technical trade-offs and experimental limitations. Super-resolution (SR) presents a promising computational approach to overcome these challenges by generating HR images from more affordable low-resolution (LR) scans, potentially improving diagnostic accuracy and efficiency without requiring additional hardware. This survey reviews recent advances in MRI SR techniques, with a focus on deep learning (DL) approaches. It examines DL-based MRI SR methods from the perspectives of computer vision, computational imaging, inverse problems, and MR physics, covering theoretical foundations, architectural designs, learning strategies, benchmark datasets, and performance metrics. We propose a systematic taxonomy to categorize these methods and present an in-depth study of both established and emerging SR techniques applicable to MRI, considering unique challenges in clinical and research contexts. We also highlight open challenges and directions that the community needs to address. Additionally, we provide a collection of essential open-access resources, tools, and tutorials, available on our GitHub: https://github.com/mkhateri/Awesome-MRI-Super-Resolution. IEEE keywords: MRI, Super-Resolution, Deep Learning, Computational Imaging, Inverse Problem, Survey.

High-resolution (HR) magnetic resonance imaging (MRI) is crucial for many clinical and research applications. However, achieving it remains costly and constrained by technical trade-offs and experimental limitations. Super-resolution (SR) presents a promising computational approach to overcome these challenges by generating HR images from more affordable low-resolution (LR) scans, potentially improving diagnostic accuracy and efficiency without requiring additional hardware. This survey reviews recent advances in MRI SR techniques, with a focus on deep learning (DL) approaches. It examines DL-based MRI SR methods from the perspectives of computer vision, computational imaging, inverse problems, and MR physics, covering theoretical foundations, architectural designs, learning strategies, benchmark datasets, and performance metrics. We propose a systematic taxonomy to categorize these methods and present an in-depth study of both established and emerging SR techniques applicable to MRI, considering unique challenges in clinical and research contexts. We also highlight open challenges and directions that the community needs to address. Additionally, we provide a collection of essential open-access resources, tools, and tutorials, available on our GitHub: https://github.com/mkhateri/Awesome-MRI-Super-Resolution. IEEE keywords: MRI, Super-Resolution, Deep Learning, Computational Imaging, Inverse Problem, Survey.

High-resolution (HR) magnetic resonance imaging (MRI) is crucial for many clinical and research applications. However, achieving it remains costly and constrained by technical trade-offs and experimental limitations. Super-resolution (SR) presents a promising computational approach to overcome these challenges by generating HR images from more affordable low-resolution (LR) scans, potentially improving diagnostic accuracy and efficiency without requiring additional hardware. This survey reviews recent advances in MRI SR techniques, with a focus on deep learning (DL) approaches. It examines DL-based MRI SR methods from the perspectives of computer vision, computational imaging, inverse problems, and MR physics, covering theoretical foundations, architectural designs, learning strategies, benchmark datasets, and performance metrics. We propose a systematic taxonomy to categorize these methods and present an in-depth study of both established and emerging SR techniques applicable to MRI, considering unique challenges in clinical and research contexts. We also highlight open challenges and directions that the community needs to address. Additionally, we provide a collection of essential open-access resources, tools, and tutorials, available on our GitHub: https://github.com/mkhateri/Awesome-MRI-Super-Resolution. IEEE keywords: MRI, Super-Resolution, Deep Learning, Computational Imaging, Inverse Problem, Survey.

Understanding electrochemical phenomena at complex liquid solid interfaces requires linking real time structural dynamics with atomic scale interfacial chemistry. Here, we integrate operando synchrotron X-ray fluorescence and diffraction with high resolution cryogenic electron and ion multi model microscopy to provide a mechanistic understanding of Pt based alloying anodes across length scales. We directly observe the initial lithiation driven formation of Li2Pt and its evolution to a stable LiPt intermetallic phase during extended cycling via a solid solution type reaction mechanism. Simultaneously, the solid electrolyte interphase transitions from an unstable carbonate rich to a stable LiF dominated composition, confirmed by cryogenic scanning transmission electron microscopy and electron energy loss spectroscopy. Crucially, cryogenic atom probe tomography reveals spatially distinct compositional regimes within the alloy anode, including lithium flux limited, heterogeneous interfacial zone and a diffusion controlled, homogeneous LiPt alloy bulk. This nanoscale compositional gradient rationalises the emergent solid solution reaction mechanism and highlights how kinetic limitations and interface dynamics govern alloy formation and electrochemical stability. Our findings demonstrate a broadly applicable correlative framework bridging operando structural dynamics with near atomic resolution interfacial chemistry, advancing the rational design of durable alloy electrodes for next generation energy storage.

Understanding electrochemical phenomena at complex liquid solid interfaces requires linking real time structural dynamics with atomic scale interfacial chemistry. Here, we integrate operando synchrotron X-ray fluorescence and diffraction with high resolution cryogenic electron and ion multi model microscopy to provide a mechanistic understanding of Pt based alloying anodes across length scales. We directly observe the initial lithiation driven formation of Li2Pt and its evolution to a stable LiPt intermetallic phase during extended cycling via a solid solution type reaction mechanism. Simultaneously, the solid electrolyte interphase transitions from an unstable carbonate rich to a stable LiF dominated composition, confirmed by cryogenic scanning transmission electron microscopy and electron energy loss spectroscopy. Crucially, cryogenic atom probe tomography reveals spatially distinct compositional regimes within the alloy anode, including lithium flux limited, heterogeneous interfacial zone and a diffusion controlled, homogeneous LiPt alloy bulk. This nanoscale compositional gradient rationalises the emergent solid solution reaction mechanism and highlights how kinetic limitations and interface dynamics govern alloy formation and electrochemical stability. Our findings demonstrate a broadly applicable correlative framework bridging operando structural dynamics with near atomic resolution interfacial chemistry, advancing the rational design of durable alloy electrodes for next generation energy storage.

We compute and track the impact of truly nonlocal-correlation effects on the quasi-particle (QP) band-structure of hexagonal boron-nitride (h-BN) systems. To that end, we start with the consistent-exchange vdW-DF-cx version [PRB 89, 035412 (2014)] of the van der Waals density functional (vdW-DF) method [JPCM 39, 390001 (2020)] for exchange-correlation (XC) functional design and enforce piece-wise linearity in the energy changes with partial charging, using the Koopmans-integer (KI) DFT framework [JCTC 19, 7079 (2023)]. Our approach and results (denoted KI-CX) extends present-standard use of KI DFT (denoted KI-PBE as it is based on the semilocal PBE [PRL 77, 3865 (1996)] XC functional) to capture, for example, the impact of the interlayer coupling on the QPs. We contrast KI-CX and KI-PBE results for the QP band-structure and compare with both $GW$ calculations and experimental observations of the (direct and indirect) QP gaps. We find that KI-CX brings improvements in the h-BN QP energy description and generally agrees with $GW$ studies.

We compute and track the impact of truly nonlocal-correlation effects on the quasi-particle (QP) band-structure of hexagonal boron-nitride (h-BN) systems. To that end, we start with the consistent-exchange vdW-DF-cx version [PRB 89, 035412 (2014)] of the van der Waals density functional (vdW-DF) method [JPCM 39, 390001 (2020)] for exchange-correlation (XC) functional design and enforce piece-wise linearity in the energy changes with partial charging, using the Koopmans-integer (KI) DFT framework [JCTC 19, 7079 (2023)]. Our approach and results (denoted KI-CX) extends present-standard use of KI DFT (denoted KI-PBE as it is based on the semilocal PBE [PRL 77, 3865 (1996)] XC functional) to capture, for example, the impact of the interlayer coupling on the QPs. We contrast KI-CX and KI-PBE results for the QP band-structure and compare with both $GW$ calculations and experimental observations of the (direct and indirect) QP gaps. We find that KI-CX brings improvements in the h-BN QP energy description and generally agrees with $GW$ studies.

High-resolution mapping of fine particulate matter (PM2.5) is a cornerstone of sustainable urbanism but remains critically hindered by the spatial sparsity of ground monitoring networks. While traditional data-driven methods attempt to bridge this gap using satellite Aerosol Optical Depth (AOD), they often suffer from severe, non-random data missingness (e.g., due to cloud cover or nighttime) and inversion biases. To overcome these limitations, this study proposes the Spatiotemporal Physics-Guided Inference Network (SPIN), a novel framework designed for inductive spatiotemporal kriging. Unlike conventional approaches, SPIN synergistically integrates domain knowledge into deep learning by explicitly modeling physical advection and diffusion processes via parallel graph kernels. Crucially, we introduce a paradigm-shifting training strategy: rather than using error-prone AOD as a direct input, we repurpose it as a spatial gradient constraint within the loss function. This allows the model to learn structural pollution patterns from satellite data while remaining robust to data voids. Validated in the highly polluted Beijing-Tianjin-Hebei and Surrounding Areas (BTHSA), SPIN achieves a new state-of-the-art with a Mean Absolute Error (MAE) of 9.52 ug/m^3, effectively generating continuous, physically plausible pollution fields even in unmonitored areas. This work provides a robust, low-cost, and all-weather solution for fine-grained environmental management.

Lithium halides with the general formula Li$_x$M$_y$X$_6$, where M indicates transition metal ions and X halide anions are very actively studied as solid-state electrolytes, because of relatively low cost, high stability and Li conductivity. The structure and properties of these halide-based solid electrolytes (HSE) can be tuned by alloying, e.g. using different halides and/or transition metals simultaneously. The large chemical space is difficult to sample by experiments, making simulations based on broadly applicable machine-learning interatomic potentials (MLIPs) a promising approach to elucidate structure-property relations, and facilitate the design of better-performing compositions. Here we focus on the Li$_3$YCl$_{6x}$Br$_{6(1-x)}$ system, for which reliable experimental data exists, and use the recently-developed PET-MAD universal MLIP to investigate the structure of the alloy, the interplay of crystalline lattice, volume and chemical composition, and its effect on Li conductivity. We find that the distribution of Cl and Br atoms is only weakly correlated, and that the primary effect of alloying is to modulate the lattice parameter -- although it can also trigger transition between different lattice symmetries. By comparing constant-volume and constant-pressure simulations, we disentangle the effect of lattice parameter and chemical composition on the conductivity, finding that the two effects compensate each other, reducing the overall dependency of conductivity on alloy composition. We also study the effect of Y-In metal substitution finding a small increase in the conductivity for the C2/m phase at 25\% In content, and an overall higher conductivity for the P$\bar{3}$m1 phase.

Conventional strategies for enhancing creep resistance often rely on grain boundary strengthening, yet this approach can inadvertently promote premature grain boundary fracture. This study presents a subtractive alloy design strategy for nickel-based directionally solidified superalloys (DS superalloy) through elimination of conventional grain boundary strengthening elements (C, B, Zr) and the strategy improves the creep performance by 60% rivaling 2nd generation single crystal superalloys. Through characterization of heat-treated and heat-exposed microstructures, we confirm the suppression of deleterious grain boundary phases. Creep tests and fracture analysis reveal a critical transition in failure mechanism: the removal of these elements shifts the fracture mode from transgranular to intergranular. Our discussion comprehensively links this microstructural engineering to the underlying creep deformation mechanisms, showing that the enhanced performance stems from stabilization of γ channel and phase interfaces within grains, as well as strengthening of grain boundaries through serration. This work establishes a novel materials design principle that decouples grain boundary strengthening from elemental additions, offering transformative potential for next-generation high-efficiency turbine blade applications.

Many living organisms can exploit quantum mechanical effects to gain distinct biological advantages. In plants, photosynthesis uses quantum coherence to achieve near 100% efficiency in energy transfer. With advances in experimental techniques, two-dimensional electronic spectroscopy can reveal dynamic processes such as coherence and coupling within a system, and it plays an important role in studying energy transfer in photosynthesis. On the theory side, methods such as the generalized Bloch-Redfield theory and the hierarchical equations of motion are used to model photosynthetic systems. Quantum simulation, as a high-efficiency and low-complexity approach, has also made progress across various platforms in the study of photosynthesis. In recent years, a series of studies has introduced quantum coherence into artificial systems to enhance energy transfer efficiency, laying the groundwork for the design of coherent devices with efficient energy transport. Birds can use the weak geomagnetic field and spin-dependent chemical reactions to detect direction. Theoretical frameworks for animal navigation include magnetite-based mechanisms, magnetoreceptor genes, and the radical-pair mechanism. Quantum simulations of navigation have also advanced on multiple platforms. Inspired by animal navigation, diverse quantum effects have been applied to improve sensing and to support navigation tasks. This paper presents a comprehensive review of progress on quantum coherence in photosynthesis and avian navigation, along with related theoretical methods, quantum simulation approaches, and research on quantum coherent devices.

Heterogeneous porous materials play a crucial role in various engineering systems. Microstructure characterization and reconstruction provide effective means for modeling these materials, which are critical for conducting physical property simulations, structure-property linkage studies, and enhancing their performance across different applications. To achieve superior controllability and applicability with small sample sizes, we propose a statistically controllable microstructure reconstruction framework that integrates neural networks with sliced-Wasserstein metric. Specifically, our approach leverages local pattern distribution for microstructure characterization and employs a controlled sampling strategy to generate target distributions that satisfy given conditional parameters. A neural network-based model establishes the mapping from the input distribution to the target local pattern distribution, enabling microstructure reconstruction. Combinations of sliced-Wasserstein metric and gradient optimization techniques minimize the distance between these distributions, leading to a stable and reliable model. Our method can perform stochastic and controllable reconstruction tasks even with small sample sizes. Additionally, it can generate large-size (e.g. 512 and 1024) 3D microstructures using a chunking strategy. By introducing spatial location masks, our method excels at generating spatially heterogeneous and complex microstructures. We conducted experiments on stochastic reconstruction, controllable reconstruction, heterogeneous reconstruction, and large-size microstructure reconstruction across various materials. Comparative analysis through visualization, statistical measures, and physical property simulations demonstrates the effectiveness, providing new insights and possibilities for research on structure-property linkage and material inverse design.

Recent advances in generative models have introduced a new paradigm for the inverse design of inorganic materials, enabling the discovery of new crystalline structures with desired properties. However, existing generative models focus solely on structural aspects of materials during generation, while overlooking the underlying electronic behavior that fundamentally governs materials' stability and functionality. In this work, we present ChargeDIFF, the first generative model for inorganic materials that explicitly incorporates electronic structure into the generation process. Specifically, ChargeDIFF leverages charge density, a direct spatial representation of a material's electronic structure, as an additional modality for generation. ChargeDIFF demonstrates exceptional performance in both unconditional and conditional generation tasks compared to baseline models, with ablation studies revealing that this improvement is directly due to its ability to capture the material's electronic structure during generation. Moreover, the ability to control charge density during generation allows ChargeDIFF to introduce a novel inverse design method based on three-dimensional charge density, illustrating the potential to generate lithium-ion battery cathode materials with desired ion migration pathways, as validated by physics-based simulations. By highlighting the importance of accounting for electronic characteristics during material generation, ChargeDIFF offers new possibilities in the generative design of stable and functional materials.

Intelligent wearable systems are at the forefront of precision medicine and play a crucial role in enhancing human-machine interaction. Traditional devices often encounter limitations due to their dependence on empirical material design and basic signal processing techniques. To overcome these issues, we introduce the concept of Human-Symbiotic Health Intelligence (HSHI), which is a framework that integrates multi-modal sensor networks with edge-cloud collaborative computing and a hybrid approach to data and knowledge modeling. HSHI is designed to adapt dynamically to both inter-individual and intra-individual variability, transitioning health management from passive monitoring to an active collaborative evolution. The framework incorporates AI-driven optimization of materials and micro-structures, provides robust interpretation of multi-modal signals, and utilizes a dual mechanism that merges population-level insights with personalized adaptations. Moreover, the integration of closed-loop optimization through reinforcement learning and digital twins facilitates customized interventions and feedback. In general, HSHI represents a significant shift in healthcare, moving towards a model that emphasizes prevention, adaptability, and a harmonious relationship between technology and health management.

Neutrino telescopes are large-scale detectors designed to observe Cherenkov radiation produced from neutrino interactions in water or ice. They exist to identify extraterrestrial neutrino sources and to probe fundamental questions pertaining to the elusive neutrino itself. A central challenge common across neutrino telescopes is to solve a series of inverse problems known as event reconstruction, which seeks to resolve properties of the incident neutrino, based on the detected Cherenkov light. In recent times, significant efforts have been made in adapting advances from deep learning research to event reconstruction, as such techniques provide several benefits over traditional methods. While a large degree of similarity in reconstruction needs and low-level data exists, cross-experimental collaboration has been hindered by a lack of diverse open-source datasets for comparing methods. We present NuBench, an open benchmark for deep learning-based event reconstruction in neutrino telescopes. NuBench comprises seven large-scale simulated datasets containing nearly 130 million charged- and neutral-current muon-neutrino interactions spanning 10 GeV to 100 TeV, generated across six detector geometries inspired by existing and proposed experiments. These datasets provide pulse- and event-level information suitable for developing and comparing machine-learning reconstruction methods in both water and ice environments. Using NuBench, we evaluate four reconstruction algorithms - ParticleNeT and DynEdge, both actively used within the KM3NeT and IceCube collaborations, respectively, along with GRIT and DeepIce - on up to five core tasks: energy and direction reconstruction, topology classification, interaction vertex prediction, and inelasticity estimation.

Neutrino telescopes are large-scale detectors designed to observe Cherenkov radiation produced from neutrino interactions in water or ice. They exist to identify extraterrestrial neutrino sources and to probe fundamental questions pertaining to the elusive neutrino itself. A central challenge common across neutrino telescopes is to solve a series of inverse problems known as event reconstruction, which seeks to resolve properties of the incident neutrino, based on the detected Cherenkov light. In recent times, significant efforts have been made in adapting advances from deep learning research to event reconstruction, as such techniques provide several benefits over traditional methods. While a large degree of similarity in reconstruction needs and low-level data exists, cross-experimental collaboration has been hindered by a lack of diverse open-source datasets for comparing methods. We present NuBench, an open benchmark for deep learning-based event reconstruction in neutrino telescopes. NuBench comprises seven large-scale simulated datasets containing nearly 130 million charged- and neutral-current muon-neutrino interactions spanning 10 GeV to 100 TeV, generated across six detector geometries inspired by existing and proposed experiments. These datasets provide pulse- and event-level information suitable for developing and comparing machine-learning reconstruction methods in both water and ice environments. Using NuBench, we evaluate four reconstruction algorithms - ParticleNeT and DynEdge, both actively used within the KM3NeT and IceCube collaborations, respectively, along with GRIT and DeepIce - on up to five core tasks: energy and direction reconstruction, topology classification, interaction vertex prediction, and inelasticity estimation.

Microbial Fuel Cells (MFCs) offer a promising pathway for sustainable energy generation by converting organic matter into electricity through microbial processes. A key factor influencing MFC performance is the anode structure, where design and material properties play a crucial role. Existing predictive models struggle to capture the complex geometric dependencies necessary to optimize these structures. To solve this problem, we propose KANGURA: Kolmogorov-Arnold Network-Based Geometry-Aware Learning with Unified Representation Attention. KANGURA introduces a new approach to three-dimensional (3D) machine learning modeling. It formulates prediction as a function decomposition problem, where Kolmogorov-Arnold Network (KAN)- based representation learning reconstructs geometric relationships without a conventional multi- layer perceptron (MLP). To refine spatial understanding, geometry-disentangled representation learning separates structural variations into interpretable components, while unified attention mechanisms dynamically enhance critical geometric regions. Experimental results demonstrate that KANGURA outperforms over 15 state-of-the-art (SOTA) models on the ModelNet40 benchmark dataset, achieving 92.7% accuracy, and excels in a real-world MFC anode structure problem with 97% accuracy. This establishes KANGURA as a robust framework for 3D geometric modeling, unlocking new possibilities for optimizing complex structures in advanced manufacturing and quality-driven engineering applications.

Understanding and continuously refining multimodal molecular knowledge is crucial for advancing biomedicine, chemistry, and materials science. Molecule language models (MoLMs) have become powerful tools in these domains, integrating structural representations (e.g., SMILES strings, molecular graphs) with rich contextual descriptions (e.g., physicochemical properties). However, MoLMs can encode and propagate inaccuracies due to outdated web-mined training corpora or malicious manipulation, jeopardizing downstream discovery pipelines. While knowledge editing has been explored for general-domain AI, its application to MoLMs remains uncharted, presenting unique challenges due to the multifaceted and interdependent nature of molecular knowledge. In this paper, we take the first step toward MoLM editing for two critical tasks: molecule-to-caption generation and caption-to-molecule generation. To address molecule-specific challenges, we propose MolEdit, a powerful framework that enables targeted modifications while preserving unrelated molecular knowledge. MolEdit combines a Multi-Expert Knowledge Adapter that routes edits to specialized experts for different molecular facets with an Expertise-Aware Editing Switcher that activates the adapters only when input closely matches the stored edits across all expertise, minimizing interference with unrelated knowledge. To systematically evaluate editing performance, we introduce MEBench, a comprehensive benchmark assessing multiple dimensions, including Reliability (accuracy of the editing), Locality (preservation of irrelevant knowledge), and Generality (robustness to reformed queries). Across extensive experiments on two popular MoLM backbones, MolEdit delivers up to 18.8% higher Reliability and 12.0% better Locality than baselines while maintaining efficiency. The code is available at: https://github.com/LzyFischer/MolEdit.

Understanding and continuously refining multimodal molecular knowledge is crucial for advancing biomedicine, chemistry, and materials science. Molecule language models (MoLMs) have become powerful tools in these domains, integrating structural representations (e.g., SMILES strings, molecular graphs) with rich contextual descriptions (e.g., physicochemical properties). However, MoLMs can encode and propagate inaccuracies due to outdated web-mined training corpora or malicious manipulation, jeopardizing downstream discovery pipelines. While knowledge editing has been explored for general-domain AI, its application to MoLMs remains uncharted, presenting unique challenges due to the multifaceted and interdependent nature of molecular knowledge. In this paper, we take the first step toward MoLM editing for two critical tasks: molecule-to-caption generation and caption-to-molecule generation. To address molecule-specific challenges, we propose MolEdit, a powerful framework that enables targeted modifications while preserving unrelated molecular knowledge. MolEdit combines a Multi-Expert Knowledge Adapter that routes edits to specialized experts for different molecular facets with an Expertise-Aware Editing Switcher that activates the adapters only when input closely matches the stored edits across all expertise, minimizing interference with unrelated knowledge. To systematically evaluate editing performance, we introduce MEBench, a comprehensive benchmark assessing multiple dimensions, including Reliability (accuracy of the editing), Locality (preservation of irrelevant knowledge), and Generality (robustness to reformed queries). Across extensive experiments on two popular MoLM backbones, MolEdit delivers up to 18.8% higher Reliability and 12.0% better Locality than baselines while maintaining efficiency. The code is available at: https://github.com/LzyFischer/MolEdit.

Assurance for artificial intelligence (AI) systems remains fragmented across software supply-chain security, adversarial machine learning, and governance documentation. Existing transparency mechanisms - including Model Cards, Datasheets, and Software Bills of Materials (SBOMs) - advance provenance reporting but rarely provide verifiable, machine-readable evidence of model security. This paper introduces the AI Risk Scanning (AIRS) Framework, a threat-model-based, evidence-generating framework designed to operationalize AI assurance. The AIRS Framework evolved through three progressive pilot studies - Smurf (AIBOM schema design), OPAL (operational validation), and Pilot C (AIRS) - that reframed AI documentation from descriptive disclosure toward measurable, evidence-bound verification. The framework aligns its assurance fields to the MITRE ATLAS adversarial ML taxonomy and automatically produces structured artifacts capturing model integrity, packaging and serialization safety, structural adapters, and runtime behaviors. Currently, the AIRS Framework is scoped to provide model-level assurances for LLMs, but it could be expanded to include other modalities and cover system-level threats (e.g. application-layer abuses, tool-calling). A proof-of-concept on a quantized GPT-OSS-20B model demonstrates enforcement of safe loader policies, per-shard hash verification, and contamination and backdoor probes executed under controlled runtime conditions. Comparative analysis with SBOM standards of SPDX 3.0 and CycloneDX 1.6 reveals alignment on identity and evaluation metadata, but identifies critical gaps in representing AI-specific assurance fields. The AIRS Framework thus extends SBOM practice to the AI domain by coupling threat modeling with automated, auditable evidence generation, providing a principled foundation for standardized, trustworthy, and machine-verifiable AI risk documentation.

The emergence of deep learning has brought the long-standing goal of comprehensively understanding and exploring crystalline materials closer to reality. Yet, universal exploration across all elements remains hindered by the combinatorial explosion of possible chemical environments, making it difficult to balance accuracy and efficiency. Crucially, within any finite set of elements, the diversity of short-range bonding types and local geometric motifs is inherently limited. Guided by this chemical intuition, we propose a chemical-system-centric strategy for crystalline materials exploration. In this framework, generative models are coupled with machine-learned force fields as fast energy evaluators, and both are iteratively refined in a closed-loop cycle of generation, evaluation, and fine-tuning. Using the Li-P-S ternary system as a case study, we show that this approach captures the diversity of local environments with minimal additional first-principles data while maintaining structural creativity, achieving closed-loop convergence toward chemical completeness within a bounded chemical space. We further demonstrate downstream applications, including phase-diagram construction, ionic-diffusivity screening, and electronic-structure prediction. Together, this strategy provides a systematic and data-efficient framework for modeling both atomistic and electronic structures within defined chemical spaces, bridging accuracy and efficiency, and paving the way toward scalable, AI-driven discovery of crystalline materials with human-level creativity and first-principles fidelity.

Deep hashing improves retrieval efficiency through compact binary codes, yet it introduces severe and often overlooked privacy risks. The ability to reconstruct original training data from hash codes could lead to serious threats such as biometric forgery and privacy breaches. However, model inversion attacks specifically targeting deep hashing models remain unexplored, leaving their security implications unexamined. This research gap stems from the inaccessibility of genuine training hash codes and the highly discrete Hamming space, which prevents existing methods from adapting to deep hashing. To address these challenges, we propose DHMI, the first diffusion-based model inversion framework designed for deep hashing. DHMI first clusters an auxiliary dataset to derive semantic hash centers as surrogate anchors. It then introduces a surrogate-guided denoising optimization method that leverages a novel attack metric (fusing classification consistency and hash proximity) to dynamically select candidate samples. A cluster of surrogate models guides the refinement of these candidates, ensuring the generation of high-fidelity and semantically consistent images. Experiments on multiple datasets demonstrate that DHMI successfully reconstructs high-resolution, high-quality images even under the most challenging black-box setting, where no training hash codes are available. Our method outperforms the existing state-of-the-art model inversion attacks in black-box scenarios, confirming both its practical efficacy and the critical privacy risks inherent in deep hashing systems.

Deep hashing improves retrieval efficiency through compact binary codes, yet it introduces severe and often overlooked privacy risks. The ability to reconstruct original training data from hash codes could lead to serious threats such as biometric forgery and privacy breaches. However, model inversion attacks specifically targeting deep hashing models remain unexplored, leaving their security implications unexamined. This research gap stems from the inaccessibility of genuine training hash codes and the highly discrete Hamming space, which prevents existing methods from adapting to deep hashing. To address these challenges, we propose DHMI, the first diffusion-based model inversion framework designed for deep hashing. DHMI first clusters an auxiliary dataset to derive semantic hash centers as surrogate anchors. It then introduces a surrogate-guided denoising optimization method that leverages a novel attack metric (fusing classification consistency and hash proximity) to dynamically select candidate samples. A cluster of surrogate models guides the refinement of these candidates, ensuring the generation of high-fidelity and semantically consistent images. Experiments on multiple datasets demonstrate that DHMI successfully reconstructs high-resolution, high-quality images even under the most challenging black-box setting, where no training hash codes are available. Our method outperforms the existing state-of-the-art model inversion attacks in black-box scenarios, confirming both its practical efficacy and the critical privacy risks inherent in deep hashing systems.

This manuscript addresses the challenge of designing functionally graded materials (FGMs) for arbitrary-shaped domains. Towards this goal, the present work proposes a generic volume fraction profile generation algorithm based on Gaussian Process Regression (GPR). The proposed algorithm can handle complex-shaped domains and generate smooth FGM profiles while adhering to the specified volume fraction values at boundaries/part of boundaries. The resulting design space from GPR comprises diverse profiles, enhancing the potential for discovering optimal configurations. Further, the algorithm allows the user to control the smoothness of the underlying profiles and the size of the design space through a length scale parameter. Further, the proposed profile generation scheme is coupled with the genetic algorithm to find the optimum FGM profiles for a given application. To make the genetic algorithm consistent with the GPR profile generation scheme, the standard simulated binary crossover operator in the genetic algorithm has been modified with a projection operator. We present numerous thermoelastic optimization examples to demonstrate the efficacy of the proposed profile generation algorithm and optimization framework.

Aligning molecular sequence representations (e.g., SMILES notations) with textual descriptions is critical for applications spanning drug discovery, materials design, and automated chemical literature analysis. Existing methodologies typically treat molecular captioning (molecule-to-text) and text-based molecular design (text-to-molecule) as separate tasks, relying on supervised fine-tuning or contrastive learning pipelines. These approaches face three key limitations: (i) conventional metrics like BLEU prioritize linguistic fluency over chemical accuracy, (ii) training datasets frequently contain chemically ambiguous narratives with incomplete specifications, and (iii) independent optimization of generation directions leads to bidirectional inconsistency. To address these issues, we propose RTMol, a bidirectional alignment framework that unifies molecular captioning and text-to-SMILES generation through self-supervised round-trip learning. The framework introduces novel round-trip evaluation metrics and enables unsupervised training for molecular captioning without requiring paired molecule-text corpora. Experiments demonstrate that RTMol enhances bidirectional alignment performance by up to 47% across various LLMs, establishing an effective paradigm for joint molecule-text understanding and generation.

Aligning molecular sequence representations (e.g., SMILES notations) with textual descriptions is critical for applications spanning drug discovery, materials design, and automated chemical literature analysis. Existing methodologies typically treat molecular captioning (molecule-to-text) and text-based molecular design (text-to-molecule) as separate tasks, relying on supervised fine-tuning or contrastive learning pipelines. These approaches face three key limitations: (i) conventional metrics like BLEU prioritize linguistic fluency over chemical accuracy, (ii) training datasets frequently contain chemically ambiguous narratives with incomplete specifications, and (iii) independent optimization of generation directions leads to bidirectional inconsistency. To address these issues, we propose RTMol, a bidirectional alignment framework that unifies molecular captioning and text-to-SMILES generation through self-supervised round-trip learning. The framework introduces novel round-trip evaluation metrics and enables unsupervised training for molecular captioning without requiring paired molecule-text corpora. Experiments demonstrate that RTMol enhances bidirectional alignment performance by up to 47% across various LLMs, establishing an effective paradigm for joint molecule-text understanding and generation.

We apply preference learning to the task of language model-guided design of novel structural alloys. In contrast to prior work that focuses on generating stable inorganic crystals, our approach targets the synthesizeability of a specific structural class: BCC/B2 superalloys, an underexplored family of materials with potential applications in extreme environments. Using three open-weight models (LLaMA-3.1, Gemma-2, and OLMo-2), we demonstrate that language models can be optimized for multiple design objectives using a single, unified reward signal through Direct Preference Optimization (DPO). Unlike prior approaches that rely on heuristic or human-in-the-loop feedback (costly), our reward signal is derived from thermodynamic phase calculations, offering a scientifically grounded criterion for model tuning. To our knowledge, this is the first demonstration of preference-tuning a language model using physics-grounded feedback for structural alloy design. The resulting framework is general and extensible, providing a path forward for intelligent design-space exploration across a range of physical science domains.

Autonomous laboratories typically rely on data-driven decision-making, occasionally with human-in-the-loop oversight to inject domain expertise. Fully leveraging AI agents, however, requires tightly coupled, collaborative workflows spanning hypothesis generation, experimental planning, execution, and interpretation. To address this, we develop and deploy a human-AI collaborative (HAIC) workflow that integrates large language models for hypothesis generation and analysis, with collaborative policy updates driving autonomous pulsed laser deposition (PLD) experiments for remote epitaxy of BaTiO$_3$/graphene. HAIC accelerated the hypothesis formation and experimental design and efficiently mapped the growth space to graphene-damage. In situ Raman spectroscopy reveals that chemistry drives degradation while the highest energy plume components seed defects, identifying a low-O$_2$ pressure low-temperature synthesis window that preserves graphene but is incompatible with optimal BaTiO$_3$ growth. Thus, we show a two-step Ar/O$_2$ deposition is required to exfoliate ferroelectric BaTiO$_3$ while maintaining a monolayer graphene interlayer. HAIC stages human insight with AI reasoning between autonomous batches to drive rapid scientific progress, providing an evolution to many existing human-in-the-loop autonomous workflows.

Interoperability and cross-validation remains a significant challenge in the computational materials discovery community. In this context, we introduce a common input/output standard designed for internal translation by various workflow managers (AiiDA, PerQueue, Pipeline Pilot, and SimStack) to produce results in a unified schema. This standard aims to enable engine-agnostic workflow execution across multiple density functional theory (DFT) codes, including CASTEP, GPAW, Quantum ESPRESSO, and VASP. As a demonstration, we have implemented a workflow to calculate the open-circuit voltage across several battery cathode materials using the proposed universal input/output schema. We analyze and resolve the challenges of reconciling energetics computed by different DFT engines and document the code-specific idiosyncrasies that make straightforward comparisons difficult. Motivated by these challenges, we outline general design principles for robust automated DFT workflows. This work represents a practical step towards more reproducible and interoperable workflows for high-throughput materials screening, while highlighting challenges of aligning electronic properties, especially for non-pristine structures.

Understanding reactor-pressure-vessel steel microstructure is crucial for predicting mechanical properties, as carbide precipitates both strengthen the alloy and can initiate cracks. In scanning electron microscopy images, gray-value overlap between carbides and matrix makes simple thresholding ineffective. We present a data-efficient segmentation pipeline using a lightweight U-Net (30.7~M parameters) trained on just \textbf{10 annotated scanning electron microscopy images}. Despite limited data, our model achieves a \textbf{Dice-Sørensen coefficient of 0.98}, significantly outperforming the state-of-the-art in the field of metallurgy (classical image analysis: 0.85), while reducing annotation effort by one order of magnitude compared to the state-of-the-art data efficient segmentation model. This approach enables rapid, automated carbide quantification for alloy design and generalizes to other steel types, demonstrating the potential of data-efficient deep learning in reactor-pressure-vessel steel analysis.

The search for materials with intrinsically low thermal conductivity ($κ_L$) is critical for energy applications, yet conventional descriptors often fail to capture the complex interplay between bonding and lattice dynamics. Here, first-principles calculations are used to contrast the thermal transport in covalent zincblende (zb) and metavalent rocksalt (rs) BeO. We find that the metavalent bonding in rs-BeO enhances lattice anharmonicity, activating multi-phonon scattering channels and suppressing phonon transport. This results in an ultralow $κ_L$ of 24 W m$^{-1}$ K$^{-1}$ at 300 K, starkly contrasting with the zb phase (357 W m$^{-1}$ K$^{-1}$). Accurately modeling such strongly anharmonic systems requires explicit inclusion of temperature-dependent phonon renormalization and four-phonon scattering. These contributions, negligible in zb-BeO, are essential for high-precision calculations of the severely suppressed $κ_L$ in rs-BeO. Finally, we identify three key indicators to guide the discovery of metavalently bonded, incipient-metallic materials: (i) an NaCl-type crystal structure, (ii) large Grüneisen parameters ($\textgreater$2), and (iii) a breakdown of the Lyddane-Sachs-Teller relation. These findings provide microscopic insight into thermal transport suppression by metavalent bonding and offer a predictive framework for identifying promising thermoelectrics and phase-change materials.

We propose a spectrally normalized surrogate for forward and inverse mechanical homogenization with hard physical guarantees. Leveraging the Voigt-Reuss bounds, we factor their difference via a Cholesky-like operator and learn a dimensionless, symmetric positive semi-definite representation with eigenvalues in $[0,1]$; the inverse map returns symmetric positive-definite predictions that lie between the bounds in the Löwner sense. In 3D linear elasticity on an open dataset of stochastic biphasic microstructures, a fully connected Voigt-Reuss net trained on $>\!7.5\times 10^{5}$ FFT-based labels with 236 isotropy-invariant descriptors and three contrast parameters recovers the isotropic projection with near-perfect fidelity (isotropy-related entries: $R^2 \ge 0.998$), while anisotropy-revealing couplings are unidentifiable from $SO(3)$-invariant inputs. Tensor-level relative Frobenius errors have median $\approx 1.7\%$ and mean $\approx 3.4\%$ across splits. For 2D plane strain on thresholded trigonometric microstructures, coupling spectral normalization with a differentiable renderer and a CNN yields $R^2>0.99$ on all components, subpercent normalized losses, accurate tracking of percolation-induced eigenvalue jumps, and robust generalization to out-of-distribution images. Treating the parametric microstructure as design variables, batched first-order optimization with a single surrogate matches target tensors within a few percent and returns diverse near-optimal designs. Overall, the Voigt-Reuss net unifies accurate, physically admissible forward prediction with large-batch, constraint-consistent inverse design, and is generic to elliptic operators and coupled-physics settings.

We propose a spectrally normalized surrogate for forward and inverse mechanical homogenization with hard physical guarantees. Leveraging the Voigt-Reuss bounds, we factor their difference via a Cholesky-like operator and learn a dimensionless, symmetric positive semi-definite representation with eigenvalues in $[0,1]$; the inverse map returns symmetric positive-definite predictions that lie between the bounds in the Löwner sense. In 3D linear elasticity on an open dataset of stochastic biphasic microstructures, a fully connected Voigt-Reuss net trained on $>\!7.5\times 10^{5}$ FFT-based labels with 236 isotropy-invariant descriptors and three contrast parameters recovers the isotropic projection with near-perfect fidelity (isotropy-related entries: $R^2 \ge 0.998$), while anisotropy-revealing couplings are unidentifiable from $SO(3)$-invariant inputs. Tensor-level relative Frobenius errors have median $\approx 1.7\%$ and mean $\approx 3.4\%$ across splits. For 2D plane strain on thresholded trigonometric microstructures, coupling spectral normalization with a differentiable renderer and a CNN yields $R^2>0.99$ on all components, subpercent normalized losses, accurate tracking of percolation-induced eigenvalue jumps, and robust generalization to out-of-distribution images. Treating the parametric microstructure as design variables, batched first-order optimization with a single surrogate matches target tensors within a few percent and returns diverse near-optimal designs. Overall, the Voigt-Reuss net unifies accurate, physically admissible forward prediction with large-batch, constraint-consistent inverse design, and is generic to elliptic operators and coupled-physics settings.

Time Projection Chambers (TPCs) are versatile detectors that reconstruct charged-particle tracks in an ionizing medium, enabling sensitive measurements across a wide range of nuclear physics experiments. We explore sparse convolutional networks for representation learning on TPC data, finding that a sparse ResNet architecture, even with randomly set weights, provides useful structured vector embeddings of events. Pre-training this architecture on a simple physics-motivated binary classification task further improves the embedding quality. Using data from the GAseous Detector with GErmanium Tagging (GADGET) II TPC, a detector optimized for measuring low-energy $β$-delayed particle decays, we represent raw pad-level signals as sparse tensors, train Minkowski Engine ResNet models, and probe the resulting event-level embeddings which reveal rich event structure. As a cross-detector test, we embed data from the Active-Target TPC (AT-TPC) -- a detector designed for nuclear reaction studies in inverse kinematics -- using the same encoder. We find that even an untrained sparse ResNet model provides useful embeddings of AT-TPC data, and we observe improvements when the model is trained on GADGET data. Together, these results highlight the potential of sparse convolutional techniques as a general tool for representation learning in diverse TPC experiments.

Understanding degradation processes in lithium ion batteries is essential for improving long term performance and advancing sustainable energy technologies. Tin selenide (SnSe) has emerged as a promising anode material due to the high theoretical capacity of tin. Unlike conventional intercalation based electrodes, SnSe undergoes conversion and alloying reactions with lithium to form Li4.4Sn, Sn, and Li2Se, enabling high lithium storage but inducing large volume changes that cause mechanical instability and capacity fading. Embedding SnSe nanoparticles within a Ti3C2Tx MXene framework offers a strategy to mitigate these effects by enhancing conductivity and structural resilience. Here, cryogenic focused ion beam (cryo FIB) slice and view revealed progressive material redistribution and morphological transformation during cycling, underscoring the need for site specific chemical analysis. Cryogenic atom probe tomography (cryo APT) of selected regions provided high spatial and chemical resolution while preserving beam sensitive phases, uncovering nanoscale degradation mechanisms including phase transformations, partial dissolution of active material, and, importantly, the first direct evidence of copper corrosion and copper ion migration from the current collector into the electrode. The observation of copper redistribution demonstrates that current collector degradation contributes directly to chemical contamination and capacity fading in composite electrodes. Together, cryo FIB and cryo APT provide a powerful workflow for elucidating electrode degradation in reactive, beam sensitive systems, offering critical insights for designing more durable and stable next generation battery materials.

Constitutive models are fundamental to solid mechanics and materials science, underpinning the quantitative description and prediction of material responses under diverse loading conditions. Traditional phenomenological models, which are derived through empirical fitting, often lack generalizability and rely heavily on expert intuition and predefined functional forms. In this work, we propose a graph-based equation discovery framework for the automated discovery of constitutive laws directly from multisource experimental data. This framework expresses equations as directed graphs, where nodes represent operators and variables, edges denote computational relations, and edge features encode parametric dependencies. This enables the generation and optimization of free-form symbolic expressions with undetermined material-specific parameters. Through the proposed framework, we have discovered new constitutive models for strain-rate effects in alloy steel materials and the deformation behavior of lithium metal. Compared with conventional empirical models, these new models exhibit compact analytical structures and achieve higher accuracy. The proposed graph-based equation discovery framework provides a generalizable and interpretable approach for data-driven scientific modelling, particularly in contexts where traditional empirical formulations are inadequate for representing complex physical phenomena.

Antiphase boundaries (APBs) are ubiquitous in ordered Heusler alloys and strongly influence magnetic coercivity in Ni-Mn-Ga, yet the link between their atomic-scale exchange interactions and micrometer-scale magnetic contrast measured by magnetic force microscopy (MFM) remains unclear. We combine density functional theory (DFT) and finite-element magnetostatics to bridge these scales in Ni-Mn-Ga. DFT calculations on supercells containing planar APBs show that the lowest-energy configuration comprises a pair of parallel APBs enclosing a nanoscale region - only three Mn-Ga atomic layers thick - whose magnetization is antiparallel to the surrounding matrix due to strong antiferromagnetic exchange across each APB (in contrast to ferromagnetic coupling in bulk martensite). According to our magnetostatic finite element model, this thin region with antiparallel magnetization generates the characteristic MFM contrast extending approx. 100 nm from the APB pair. When the APBs are further apart than 50 nm, dipole-dipole penalties outweigh exchange gains, preventing formation of an extended antiparallel domain, in agreement with experimental evidence. These results identify APB pairs as the origin of the observed MFM contrast and offer an interpretation of the modest strengths of domain-wall pinning by APBs, informing the design of magnetic shape-memory alloys with tailored coercivity.

Microscopy such as Scanning Tunneling Microscopy (STM), Atomic Force Microscopy (AFM) and Scanning Electron Microscopy (SEM) are essential tools in material imaging at micro- and nanoscale resolutions to extract physical knowledge and materials structure-property relationships. However, tuning microscopy controls (e.g. scanning speed, current setpoint, tip bias etc.) to obtain a high-quality of images is a non-trivial and time-consuming effort. On the other hand, with sub-standard images, the key features are not accurately discovered due to noise and artifacts, leading to erroneous analysis. Existing denoising models mostly build on generalizing the weak signals as noises while the strong signals are enhanced as key features, which is not always the case in microscopy images, thus can completely erase a significant amount of hidden physical information. To address these limitations, we propose a global denoising model (GDM) to smartly remove artifacts of microscopy images while preserving weaker but physically important features. The proposed model is developed based on 1) first designing a two-imaging input channel of non-pair and goal specific pre-processed images with user-defined trade-off information between two channels and 2) then integrating a loss function of pixel- and fast Fourier-transformed (FFT) based on training the U-net model. We compared the proposed GDM with the non-FFT denoising model over STM-generated images of Copper(Cu) and Silicon(Si) materials, AFM-generated Pantoea sp.YR343 bio-film images and SEM-generated plastic degradation images. We believe this proposed workflow can be extended to improve other microscopy image quality and will benefit the experimentalists with the proposed design flexibility to smartly tune via domain-experts preferences.

Coupled ionic and electronic transport underpins processes as diverse as electrochemical energy conversion, biological signaling, and soft adaptive electronics. Yet, how chemical environments such as pH modulate this coupling at the molecular scale remains poorly understood. Here, we show that the protonation state of carboxylated polythiophenes provides precise chemical control over ion dynamics, doping efficiency, solvent uptake and mechanical response. The findings establish molecular acidity as a general strategy to program ionic preference and mechanical stability, offering design principles for pH-responsive mixed conductors and soft electronic materials that couple ionic, electronic, and mechanical functionality.

The rise of 3D generative models has enabled automatic 3D geometry and texture synthesis from multimodal inputs (e.g., text or images). However, these methods often ignore physical constraints and manufacturability considerations. In this work, we address the challenge of producing 3D designs that are both lightweight and self-supporting. We present DensiCrafter, a framework for generating lightweight, self-supporting 3D hollow structures by optimizing the density field. Starting from coarse voxel grids produced by Trellis, we interpret these as continuous density fields to optimize and introduce three differentiable, physically constrained, and simulation-free loss terms. Additionally, a mass regularization penalizes unnecessary material, while a restricted optimization domain preserves the outer surface. Our method seamlessly integrates with pretrained Trellis-based models (e.g., Trellis, DSO) without any architectural changes. In extensive evaluations, we achieve up to 43% reduction in material mass on the text-to-3D task. Compared to state-of-the-art baselines, our method could improve the stability and maintain high geometric fidelity. Real-world 3D-printing experiments confirm that our hollow designs can be reliably fabricated and could be self-supporting.

The rise of 3D generative models has enabled automatic 3D geometry and texture synthesis from multimodal inputs (e.g., text or images). However, these methods often ignore physical constraints and manufacturability considerations. In this work, we address the challenge of producing 3D designs that are both lightweight and self-supporting. We present DensiCrafter, a framework for generating lightweight, self-supporting 3D hollow structures by optimizing the density field. Starting from coarse voxel grids produced by Trellis, we interpret these as continuous density fields to optimize and introduce three differentiable, physically constrained, and simulation-free loss terms. Additionally, a mass regularization penalizes unnecessary material, while a restricted optimization domain preserves the outer surface. Our method seamlessly integrates with pretrained Trellis-based models (e.g., Trellis, DSO) without any architectural changes. In extensive evaluations, we achieve up to 43% reduction in material mass on the text-to-3D task. Compared to state-of-the-art baselines, our method could improve the stability and maintain high geometric fidelity. Real-world 3D-printing experiments confirm that our hollow designs can be reliably fabricated and could be self-supporting.

We present a forward prediction material screening framework designed to discover Pd-Cu alloys with improved B2 phase stability, thereby unlocking simultaneous $H_2$ generation and utilization. First, we trained CatBoost models with literature-derived Pd alloy data to predict $H_2$ permeability from composition and testing conditions. We evaluated fractional, composition-based, and physics-informed descriptors, individually and in combination, and showed that sequential Pearson filtering and fold-wise SHAP-based recursive feature elimination with cross-fold aggregation reduced errors while controlling complexity. Guided by the one-SE rule, a narrower domain-informed set of 13 features provided the best accuracy parsimony trade-off ($R^2=0.81$), only 0.01 below the max. $R^2$ achievable with 3x the number of features. SHAP analysis indicated that high permeability is promoted by elevated temperature, lattice expansion relative to Pd, atomic size mismatch, and favorable mixing tendencies. Second, the selected model was applied to screen $Pd_{(100-x-y)}Cu_{x}M_{y}$ spanning 16 co-dopants M for B2 stabilization. For each M system, we obtained the Pareto set of compositions that minimize Pd content and Miedema heat of formation and maximize the permeability, then picked three compounds, including that with the highest predicted permeability, the lowest Miedema heat of formation, and the lowest Pd content. With a final filter considering M concentration for single-phase Pd-M solution formation, we recommend Pd48.48Cu43.00Y8.52, Pd49.08Cu42.45Sc8.47, Pd56.09Cu33.70La10.21, and Pd52.68Cu40.44Mg6.88 for experimental validation. We predict those alloys to exhibit permeabilities 1.7 to 1.9 higher than B2 Pd60Cu40. Our framework provides plausible experimental targets and a scalable pathway for designing stable, high-temperature, H2-selective Pd-alloy membranes.

Integration of organic semiconductors into flexible electronics requires that their optoelectronic properties remain stable under mechanical deformation. Among these, the optical band gap governs exciton generation and limits photovoltaic voltage, making it a key parameter for strain-resilient design. In this work, we investigate band gap shifts in poly(3-hexylthiophene-2,5-diyl) (P3HT) and poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS)/P3HT thin films deposited on flexible poly(ethylene terephthalate) (PET) substrates under uniaxial tensile strain ranging from 1\% to 10\%. Samples were subjected to mechanical deformation and then characterized by ultraviolet--visible (UV--Vis) absorption spectroscopy. The optical band gaps extracted using a standardized Tauc analysis and statistically validated through equivalence testing and robust regression models. We find that up to 7\% strain, the band gap shift ($ΔE_g$) remains effectively invariant, independent of annealing condition or stack configuration, demonstrating electronic stability. However, at 10\% strain, all groups exhibit a reproducible widening of $\sim$4--5~meV. This threshold-like behavior marks a transition from mechanical accommodation to electronic perturbation. These findings confirm that the optical band gap in semicrystalline P3HT-based thin films is robust under practical deformation, which provides clear strain thresholds to inform mechanical modeling and device-level simulation of flexible organic optoelectronic systems.

Diffusion models serve as a powerful generative framework for solving inverse problems. However, they still face two key challenges: 1) the distortion-perception tradeoff, where improving perceptual quality often degrades reconstruction fidelity, and 2) the exposure bias problem, where the training-inference input mismatch leads to prediction error accumulation and reduced reconstruction quality. In this work, we propose the Regularized Schrödinger Bridge (RSB), an adaptation of Schrödinger Bridge tailored for inverse problems that addresses the above limitations. RSB employs a novel regularized training strategy that perturbs both the input states and targets, effectively mitigating exposure bias by exposing the model to simulated prediction errors and also alleviating distortion by well-designed interpolation via the posterior mean. Extensive experiments on two typical inverse problems for speech enhancement demonstrate that RSB outperforms state-of-the-art methods, significantly improving distortion metrics and effectively reducing exposure bias.

Diffusion models serve as a powerful generative framework for solving inverse problems. However, they still face two key challenges: 1) the distortion-perception tradeoff, where improving perceptual quality often degrades reconstruction fidelity, and 2) the exposure bias problem, where the training-inference input mismatch leads to prediction error accumulation and reduced reconstruction quality. In this work, we propose the Regularized Schrödinger Bridge (RSB), an adaptation of Schrödinger Bridge tailored for inverse problems that addresses the above limitations. RSB employs a novel regularized training strategy that perturbs both the input states and targets, effectively mitigating exposure bias by exposing the model to simulated prediction errors and also alleviating distortion by well-designed interpolation via the posterior mean. Extensive experiments on two typical inverse problems for speech enhancement demonstrate that RSB outperforms state-of-the-art methods, significantly improving distortion metrics and effectively reducing exposure bias.

Diffusion models serve as a powerful generative framework for solving inverse problems. However, they still face two key challenges: 1) the distortion-perception tradeoff, where improving perceptual quality often degrades reconstruction fidelity, and 2) the exposure bias problem, where the training-inference input mismatch leads to prediction error accumulation and reduced reconstruction quality. In this work, we propose the Regularized Schrödinger Bridge (RSB), an adaptation of Schrödinger Bridge tailored for inverse problems that addresses the above limitations. RSB employs a novel regularized training strategy that perturbs both the input states and targets, effectively mitigating exposure bias by exposing the model to simulated prediction errors and also alleviating distortion by well-designed interpolation via the posterior mean. Extensive experiments on two typical inverse problems for speech enhancement demonstrate that RSB outperforms state-of-the-art methods, significantly improving distortion metrics and effectively reducing exposure bias.

Simulating microstructure evolution (MicroEvo) is vital for materials design but demands high numerical accuracy, efficiency, and physical fidelity. Although recent studies on deep learning (DL) offer a promising alternative to traditional solvers, the field lacks standardized benchmarks. Existing studies are flawed due to a lack of comparing specialized MicroEvo DL models with state-of-the-art spatio-temporal architectures, an overemphasis on numerical accuracy over physical fidelity, and a failure to analyze error propagation over time. To address these gaps, we introduce MicroEvoEval, the first comprehensive benchmark for image-based microstructure evolution prediction. We evaluate 14 models, encompassing both domain-specific and general-purpose architectures, across four representative MicroEvo tasks with datasets specifically structured for both short- and long-term assessment. Our multi-faceted evaluation framework goes beyond numerical accuracy and computational cost, incorporating a curated set of structure-preserving metrics to assess physical fidelity. Our extensive evaluations yield several key insights. Notably, we find that modern architectures (e.g., VMamba), not only achieve superior long-term stability and physical fidelity but also operate with an order-of-magnitude greater computational efficiency. The results highlight the necessity of holistic evaluation and identify these modern architectures as a highly promising direction for developing efficient and reliable surrogate models in data-driven materials science.

Simulating microstructure evolution (MicroEvo) is vital for materials design but demands high numerical accuracy, efficiency, and physical fidelity. Although recent studies on deep learning (DL) offer a promising alternative to traditional solvers, the field lacks standardized benchmarks. Existing studies are flawed due to a lack of comparing specialized MicroEvo DL models with state-of-the-art spatio-temporal architectures, an overemphasis on numerical accuracy over physical fidelity, and a failure to analyze error propagation over time. To address these gaps, we introduce MicroEvoEval, the first comprehensive benchmark for image-based microstructure evolution prediction. We evaluate 14 models, encompassing both domain-specific and general-purpose architectures, across four representative MicroEvo tasks with datasets specifically structured for both short- and long-term assessment. Our multi-faceted evaluation framework goes beyond numerical accuracy and computational cost, incorporating a curated set of structure-preserving metrics to assess physical fidelity. Our extensive evaluations yield several key insights. Notably, we find that modern architectures (e.g., VMamba), not only achieve superior long-term stability and physical fidelity but also operate with an order-of-magnitude greater computational efficiency. The results highlight the necessity of holistic evaluation and identify these modern architectures as a highly promising direction for developing efficient and reliable surrogate models in data-driven materials science.

The design of intelligent materials often draws parallels with the complex adaptive behaviors of biological organisms, where robust functionality stems from sophisticated hierarchical organization and emergent long-distance coordination among a myriad local components. Current synthetic materials, despite integrating advanced sensors and actuators, predominantly demonstrate only simple, pre-programmed stimulus-response functionalities, falling short of robustly autonomous intelligent behavior. These systems typically execute tasks determined by rigid design or external control, fundamentally lacking the intricate internal feedback loops, dynamic adaptation, self-generated learning, and genuine self-determination characteristic of biological agents. This perspective proposes a fundamentally different approach focusing on architectures where material-based intelligence is not pre-designed, but arises spontaneously from self-organization harnessing far-from-equilibrium dynamics. This work explores interdisciplinary concepts from material physics, chemistry, biology, and computation, identifying concrete pathways toward developing materials that not only react, but actively perceive, adapt, learn, self-correct, and potentially self-construct, moving beyond biomimicry to cultivate fully synthetic, self-evolving systems without external control. This framework outlines the fundamental requirements for, and constraints upon, future architectures where complex, goal-directed functionalities emerge synergistically from integrated local processes, distinguishing material-based intelligence from traditional hardware-software divisions. This demands that concepts of high-level goals and robust, replicable memory mechanisms are encoded and enacted through the material's inherent dynamics, inherently blurring the distinction between system output and process.

Semantic communications mark a paradigm shift from bit-accurate transmission toward meaning-centric communication, essential as wireless systems approach theoretical capacity limits. The emergence of generative AI has catalyzed generative semantic communications, where receivers reconstruct content from minimal semantic cues by leveraging learned priors. Among generative approaches, diffusion models stand out for their superior generation quality, stable training dynamics, and rigorous theoretical foundations. However, the field currently lacks systematic guidance connecting diffusion techniques to communication system design, forcing researchers to navigate disparate literatures. This article provides the first comprehensive tutorial on diffusion models for generative semantic communications. We present score-based diffusion foundations and systematically review three technical pillars: conditional diffusion for controllable generation, efficient diffusion for accelerated inference, and generalized diffusion for cross-domain adaptation. In addition, we introduce an inverse problem perspective that reformulates semantic decoding as posterior inference, bridging semantic communications with computational imaging. Through analysis of human-centric, machine-centric, and agent-centric scenarios, we illustrate how diffusion models enable extreme compression while maintaining semantic fidelity and robustness. By bridging generative AI innovations with communication system design, this article aims to establish diffusion models as foundational components of next-generation wireless networks and beyond.

Point defects govern many important functional properties of two-dimensional (2D) materials. However, resolving the three-dimensional (3D) arrangement of these defects in multi-layer 2D materials remains a fundamental challenge, hindering rational defect engineering. Here, we overcome this limitation using an artificial intelligence-guided electron microscopy workflow to map the 3D topology and clustering of atomic vacancies in Ti$_3$C$_2$T$_X$ MXene. Our approach reconstructs the 3D coordinates of vacancies across hundreds of thousands of lattice sites, generating robust statistical insight into their distribution that can be correlated with specific synthesis pathways. This large-scale data enables us to classify a hierarchy of defect structures--from isolated vacancies to nanopores--revealing their preferred formation and interaction mechanisms, as corroborated by molecular dynamics simulations. This work provides a generalizable framework for understanding and ultimately controlling point defects across large volumes, paving the way for the rational design of defect-engineered functional 2D materials.

Gaussian policies have dominated continuous control in deep reinforcement learning (RL), yet they suffer from a fundamental mismatch: their unbounded support requires ad-hoc squashing functions that distort the geometry of bounded action spaces. While von Mises-Fisher (vMF) distributions offer a theoretically grounded alternative on the sphere, their reliance on Bessel functions and rejection sampling hinders practical adoption. We propose \textbf{Geometric Action Control (GAC)}, a novel action generation paradigm that preserves the geometric benefits of spherical distributions while \textit{simplifying computation}. GAC decomposes action generation into a direction vector and a learnable concentration parameter, enabling efficient interpolation between deterministic actions and uniform spherical noise. This design reduces parameter count from \(2d\) to \(d+1\), and avoids the \(O(dk)\) complexity of vMF rejection sampling, achieving simple \(O(d)\) operations. Empirically, GAC consistently matches or exceeds state-of-the-art methods across six MuJoCo benchmarks, achieving 37.6\% improvement over SAC on Ant-v4 and the best results on 4 out of 6 tasks. Our ablation studies reveal that both \textbf{spherical normalization} and \textbf{adaptive concentration control} are essential to GAC's success. These findings suggest that robust and efficient continuous control does not require complex distributions, but a principled respect for the geometry of action spaces. Code and pretrained models are available in supplementary materials.

Standard stochastic policies for continuous control often rely on ad-hoc boundary-enforcing transformations (e.g., tanh) which can distort the underlying optimization landscape and introduce gradient pathologies. While alternative parameterizations on the unit manifold (e.g., directional distributions) are theoretically appealing, their computational complexity (often requiring special functions or rejection sampling) has limited their practical use. We propose a novel, computationally efficient action generation paradigm that preserves the structural benefits of operating on a unit manifold. Our method decomposes the action into a deterministic directional vector and a learnable concentration scalar, enabling efficient interpolation between the target direction and uniform noise on the unit manifold. This design can reduce policy head parameters by nearly 50\% (from $2d$ to $d+1$) and maintains a simple $O(d)$ sampling complexity, avoiding costly sampling procedures. Empirically, our method matches or exceeds state-of-the-art methods on standard continuous control benchmarks, with significant improvements (e.g., +37.6\% and +112\%) on high-dimensional locomotion tasks. Ablation studies confirm that both the unit-norm normalization and the adaptive concentration mechanism are essential to the method's success. These findings suggest that robust, efficient control can be achieved by explicitly respecting the structure of bounded action spaces, rather than relying on complex, unbounded distributions. Code is available in supplementary materials.

Frequency Modulated Continuous Wave (FMCW) radars can measure subtle chest wall oscillations to enable non-contact heartbeat sensing. However, traditional radar-based heartbeat sensing methods face performance degradation due to noise. Learning-based radar methods achieve better noise robustness but require costly labeled signals for supervised training. To overcome these limitations, we propose the first unsupervised framework for radar-based heartbeat sensing via Augmented Pseudo-Label and Noise Contrast (Radar-APLANC). We propose to use both the heartbeat range and noise range within the radar range matrix to construct the positive and negative samples, respectively, for improved noise robustness. Our Noise-Contrastive Triplet (NCT) loss only utilizes positive samples, negative samples, and pseudo-label signals generated by the traditional radar method, thereby avoiding dependence on expensive ground-truth physiological signals. We further design a pseudo-label augmentation approach featuring adaptive noise-aware label selection to improve pseudo-label signal quality. Extensive experiments on the Equipleth dataset and our collected radar dataset demonstrate that our unsupervised method achieves performance comparable to state-of-the-art supervised methods. Our code, dataset, and supplementary materials can be accessed from https://github.com/RadarHRSensing/Radar-APLANC.

Discovering functional crystalline materials entails navigating an immense combinatorial design space. While recent advances in generative artificial intelligence have enabled the sampling of chemically plausible compositions and structures, a fundamental challenge remains: the objective misalignment between likelihood-based sampling in generative modelling and targeted focus on underexplored regions where novel compounds reside. Here, we introduce a reinforcement learning framework that guides latent denoising diffusion models toward diverse and novel, yet thermodynamically viable crystalline compounds. Our approach integrates group relative policy optimisation with verifiable, multi-objective rewards that jointly balance creativity, stability, and diversity. Beyond de novo generation, we demonstrate enhanced property-guided design that preserves chemical validity, while targeting desired functional properties. This approach establishes a modular foundation for controllable AI-driven inverse design that addresses the novelty-validity trade-off across scientific discovery applications of generative models.

Efficiently and accurately determining the symmetry is a crucial step in the structural analysis of crystalline materials. Existing methods usually mindlessly apply deep learning models while ignoring the underlying chemical rules. More importantly, experiments show that they face a serious sub-property confusion SPC problem. To address the above challenges, from a decoupled perspective, we introduce the XRDecoupler framework, a problem-solving arsenal specifically designed to tackle the SPC problem. Imitating the thinking process of chemists, we innovatively incorporate multidimensional crystal symmetry information as superclass guidance to ensure that the model's prediction process aligns with chemical intuition. We further design a hierarchical PXRD pattern learning model and a multi-objective optimization approach to achieve high-quality representation and balanced optimization. Comprehensive evaluations on three mainstream databases (e.g., CCDC, CoREMOF, and InorganicData) demonstrate that XRDecoupler excels in performance, interpretability, and generalization.

Monte Carlo random walk methods are widely used in capacitance extraction for their mesh-free formulation and inherent parallelism. However, modern semiconductor technologies with densely packed structures present significant challenges in unbiasedly sampling transition domains in walk steps with multiple high-contrast dielectric materials. We present DeepRWCap, a machine learning-guided random walk solver that predicts the transition quantities required to guide each step of the walk. These include Poisson kernels, gradient kernels, signs and magnitudes of weights. DeepRWCap employs a two-stage neural architecture that decomposes structured outputs into face-wise distributions and spatial kernels on cube faces. It uses 3D convolutional networks to capture volumetric dielectric interactions and 2D depthwise separable convolutions to model localized kernel behavior. The design incorporates grid-based positional encodings and structural design choices informed by cube symmetries to reduce learning redundancy and improve generalization. Trained on 100,000 procedurally generated dielectric configurations, DeepRWCap achieves a mean relative error of $1.24\pm0.53$\% when benchmarked against the commercial Raphael solver on the self-capacitance estimation of 10 industrial designs spanning 12 to 55 nm nodes. Compared to the state-of-the-art stochastic difference method Microwalk, DeepRWCap achieves an average 23\% speedup. On complex designs with runtimes over 10 s, it reaches an average 49\% acceleration.

Monte Carlo random walk methods are widely used in capacitance extraction for their mesh free formulation and inherent parallelism. However, modern semiconductor technologies with densely packed structures present significant challenges in unbiasedly sampling transition domains in walk steps with multiple high contrast dielectric materials. We present DeepRWCap, a machine learning guided random walk solver that predicts the transition quantities required to guide each step of the walk. These include Poisson kernels, gradient kernels, and the signs and magnitudes of weights. DeepRWCap employs a two stage neural architecture that decomposes structured outputs into face wise distributions and spatial kernels on cube faces. It uses 3D convolutional networks to capture volumetric dielectric interactions and 2D depthwise separable convolutions to model localized kernel behavior. The design incorporates grid based positional encodings and structural design choices informed by cube symmetries to reduce learning redundancy and improve generalization. Trained on 100000 procedurally generated dielectric configurations, DeepRWCap achieves a mean relative error of 1.24 +/- 0.53% when benchmarked against the commercial Raphael solver on the self capacitance estimation of 10 industrial designs spanning 12 to 55 nm nodes. Compared to the state of the art stochastic difference method Microwalk, DeepRWCap achieves an average speedup of 23%. On complex designs with runtimes over 10 seconds, it reaches an average acceleration of 49%.

Lost circulation remains a major and costly challenge in drilling operations, often resulting in wellbore instability, stuck pipe, and extended non-productive time. Accurate prediction of fluid loss is therefore essential for improving drilling safety and efficiency. This study presents a probabilistic machine learning framework based on Gaussian Process Regression (GPR) for predicting drilling fluid loss in complex formations. The GPR model captures nonlinear dependencies among drilling parameters while quantifying predictive uncertainty, offering enhanced reliability for high-risk decision-making. Model hyperparameters are optimized using the Limited memory Broyden Fletcher Goldfarb Shanno (LBFGS) algorithm to ensure numerical stability and robust generalization. To improve interpretability, Local Interpretable Model agnostic Explanations (LIME) are employed to elucidate how individual features influence model predictions. The results highlight the potential of explainable probabilistic learning for proactive identification of lost-circulation risks, optimized design of lost circulation materials (LCM), and reduction of operational uncertainties in drilling applications.

The Wiedemann-Franz (WF) law, relating the electronic thermal conductivity ($κ_{\rm e}$) to the electrical conductivity, is vital in numerous applications such as in the design of thermoelectric materials and in the experimental determination of the lattice thermal conductivity ($κ_{\rm L}$). While the WF law is generally robust, violations are frequently observed, typically manifesting in a reduced Lorenz number ($L$) relative to the Sommerfeld value ($L_0$) due to inelastic scattering. Here, we report a pronounced departure from the WF law in the topological semimetal CoSi, where the electronic Lorenz number ($L_{\rm e}$) instead rises up to $\sim40\%$ above $L_0$. We demonstrate that this anomaly arises from strong bipolar diffusive transport, enabled by topological band-induced electron-hole compensation, which allows electrons and holes to flow cooperatively and additively enhance the heat current. Concurrently, we unveil that the lattice contribution to thermal conductivity is anomalously large and becomes the dominant component below room temperature. As a result, if $κ_{\rm L}$ is assumed negligible -- as conventional in metals, the resulting $L$ from the total thermal conductivity ($κ_{\rm tot}=κ_{\rm L}+κ_{\rm e}$) deviates from $L_0$ by more than a factor of three. Our work provides deeper insight into the unconventional thermal transport physics in topological semimetals.

An analytical solution to the nonlinear Poisson equation governing the inversion layer in metal-oxide-semiconductor (MOS) structures has recently been obtained, resolving a fundamental challenge in semiconductor theory first identified in 1955. This breakthrough enables the derivation of explicit expressions for relevant physical quantities, such as the inversion-layer width, electric potential, and charge distribution, as functions of gate voltage $V_G$, distance from oxide-semiconductor interface and impurity concentration. These quantities exhibit rapid variation during early-stage inversion but saturate once the gate voltage exceeds the threshold voltage by a few tenths of a volt signaling a transition in the MOS response to $V_G$. The onset of tunneling through the Esaki barrier leads to increased charge accumulation near the interface, reshaping the charge distribution into a two-dimensional profile and shifting the potential drop from the semiconductor to the oxide layer. This reconfiguration resembles the behavior of an ideal parallel-plate capacitor, with charge confined at the interface and the voltage drop localized across the oxide. We analyze this mechanism in detail and demonstrate, through explicit calculations, that the tunneling current through the Esaki-like barrier formed during inversion becomes dominant, effectively superseding classical inversion behavior. These results offer a new analytical foundation for quantum-aware device modeling and inform the design of next-generation MOSFET and tunneling FET architectures.

Data scarcity remains a central challenge in materials discovery, where finding meaningful descriptors and tuning models for generalization is critical but inherently a discrete optimization problem prone to multiple local minima confounding the true optimal state. Classical methods often get trapped in these minima, while quantum annealing can escape them via quantum fluctuations, including tunneling, that overcome narrow energy barriers. We present a quantum-assisted machine-learning (QaML) framework that employs quantum annealing to address these combinatorial optimization challenges through feature selection, support-vector training formulated in QUBO form for classification and regression, and a new QUBO-based neural-network pruning formulation. Recursive batching enables quantum annealing to handle large feature spaces beyond current qubit limits, while quantum-pruned networks exhibit superior generalization over classical methods, suggesting that quantum annealing preferentially samples flatter, more stable regions of the loss landscape. Applied to high-entropy alloys (HEAs), a data-limited but compositionally complex testbed, the framework integrates models for fracture-strain classification and yield-strength regression under physics-based constraints. The framework identified and experimentally validated Al8Cr38Fe50Mn2Ti2 (at.%), a single-phase BCC alloy exhibiting a 0.2 % yield strength of 568 MPa, greater than 40 % compressive strain without fracture, and a critical current density in reducing acid nearly an order of magnitude lower than 304 stainless steel. These results establish QA as a practical route to overcome classical optimization limits and accelerate materials discovery.

Graph network-based simulators (GNS) have demonstrated strong potential for learning particle-based physics (such as fluids, deformable solids, and granular flows) while generalizing to unseen geometries due to their inherent inductive biases. However, existing models are typically trained for a single material type and fail to generalize across distinct constitutive behaviors, limiting their applicability in real-world engineering settings. Using granular flows as a running example, we propose a parameter-efficient conditioning mechanism that makes the GNS model adaptive to material parameters. We identify that sensitivity to material properties is concentrated in the early message-passing (MP) layers, a finding we link to the local nature of constitutive models (e.g., Mohr-Coulomb) and their effects on information propagation. We empirically validate this by showing that fine-tuning only the first few (1-5) of 10 MP layers of a pretrained model achieves comparable test performance as compared to fine-tuning the entire network. Building on this insight, we propose a parameter-efficient Feature-wise Linear Modulation (FiLM) conditioning mechanism designed to specifically target these early layers. This approach produces accurate long-term rollouts on unseen, interpolated, or moderately extrapolated values (e.g., up to 2.5 degrees for friction angle and 0.25 kPa for cohesion) when trained exclusively on as few as 12 short simulation trajectories from new materials, representing a 5-fold data reduction compared to a baseline multi-task learning method. Finally, we validate the model's utility by applying it to an inverse problem, successfully identifying unknown cohesion parameters from trajectory data. This approach enables the use of GNS in inverse design and closed-loop control tasks where material properties are treated as design variables.

We present a conditional diffusion model for electromagnetic inverse design that generates structured media geometries directly from target differential scattering cross-section profiles, bypassing expensive iterative optimization. Our 1D U-Net architecture with Feature-wise Linear Modulation learns to map desired angular scattering patterns to 2x2 dielectric sphere structure, naturally handling the non-uniqueness of inverse problems by sampling diverse valid designs. Trained on 11,000 simulated metasurfaces, the model achieves median MPE below 19% on unseen targets (best: 1.39%), outperforming CMA-ES evolutionary optimization while reducing design time from hours to seconds. These results demonstrate that employing diffusion models is promising for advancing electromagnetic inverse design research, potentially enabling rapid exploration of complex metasurface architectures and accelerating the development of next-generation photonic and wireless communication systems. The code is publicly available at https://github.com/mikzuker/inverse_design_metasurface_generation.

Predicting quantum wavefunction probability distributions is crucial for computational chemistry and materials science, yet machine learning (ML) models often face a trade-off between accuracy and interpretability. This study compares Artificial Neural Networks (ANNs) and Adaptive Neuro-Fuzzy Inference Systems (ANFIS) in modeling quantum probability distributions for the H$_{2}^+$ ion, leveraging data generated via Physics-Informed Neural Networks (PINNs). While ANN achieved superior accuracy (R$^2$ = 0.99 vs ANFIS's 0.95 with Gaussian membership functions), it required over 50x more parameters (2,305 vs 39-45). ANFIS, however, provided unique interpretability: its Gaussian membership functions encoded spatial electron localization near proton positions ($μ= 1.2 A$), mirroring Born probability densities, while fuzzy rules reflected quantum superposition principles. Rules prioritizing the internuclear direction revealed the system's 1D symmetry, aligning with Linear Combination of Atomic Orbitals theory--a novel data-driven perspective on orbital hybridization. Membership function variances ($σ$) further quantified electron delocalization trends, and peak prediction errors highlighted unresolved quantum cusps. The choice of functions critically impacted performance: Gaussian/Generalized Bell outperformed Sigmoid, with errors improving as training data increased, showing scalability. This study underscores the context-dependent value of ML: ANN for precision and ANFIS for interpretable, parameter-efficient approximations that link inputs to physical behavior. These findings advocate hybrid approaches in quantum simulations, balancing accuracy with explainability to accelerate discovery. Future work should extend ANFIS to multi-electron systems and integrate domain-specific constraints (e.g., kinetic energy terms), bridging data-driven models and fundamental physics.

Machine learning potentials (MLPs) have significantly advanced global crystal structure prediction by enabling efficient and accurate property evaluations. In this study, global structure searches are performed for 11 bismuth-based binary systems, including Na-Bi, Ca-Bi, and Eu-Bi, under pressures ranging from 0 to 20 GPa, employing polynomial MLPs developed specifically for these systems. The searches reveal numerous compounds not previously reported in the literature and identify all experimentally known compounds that are representable within the explored configurational space. These results highlight the robustness and reliability of the current MLP-based structure search. The study provides valuable insights into the discovery and design of novel bismuth-based materials under both ambient and high-pressure conditions.

We reveal intrinsic fracture nonreciprocity, manifesting as directional asymmetry in crack resistance, in two-dimensional heterostructures engineered through lattice-mismatched interfaces. Density-functional theory combined with machine-learning molecular dynamics show that intrinsic lattice mismatch between bonded component crystals imprints asymmetric prestrain states at crack tips, governing bond-breaking thresholds through charge redistribution. The failure criterion obeys a universal exponential scaling law between normalized charge density and bond strain, insensitive to bonding chemistry and local atomic environment. The magnitude of nonreciprocity scales systematically with lattice mismatch, reaching 49% at 10% mismatch. Validation across hexagonal, square, rectangular, and oblique two-dimensional lattices confirms universality, establishing interface strain engineering as a general design principle that bridges electronic structure to nanoscale failure, enabling rational design of damage-tolerant nanostructures.

Inorganic crystal materials have broad application potential due to excellent physical and chemical properties, with elastic properties (shear modulus, bulk modulus) crucial for predicting materials' electrical conductivity, thermal conductivity and mechanical properties. Traditional experimental measurement suffers from high cost and low efficiency, while theoretical simulation and graph neural network-based machine learning methods--especially crystal graph convolutional neural networks (CGCNNs)--have become effective alternatives, achieving remarkable results in predicting material elastic properties. This study trained two CGCNN models using shear modulus and bulk modulus data of 10987 materials from the Matbench v0.1 dataset, which exhibit high accuracy (mean absolute error <13, coefficient of determination R-squared close to 1) and good generalization ability. Materials were screened to retain those with band gaps between 0.1-3.0 eV and exclude radioactive element-containing compounds. The final predicted dataset comprises two parts: 54359 crystal structures from the Materials Project database and 26305 crystal structures discovered by Merchant et al. (2023 Nature 624 80). Ultimately, this study completed the prediction of shear modulus and bulk modulus for 80664 inorganic crystals. This work enriches existing material elastic data resources and provides robust support for material design, with all data openly available at https://doi.org/10.57760/sciencedb.j00213.00104.

Topological materials--including insulators (TIs) and semimetals (TSMs)--hold immense promise for quantum technologies, yet their discovery remains constrained by the high computational cost of first-principles calculations and the slow, resource-intensive nature of experimental synthesis. Here, we introduce TXL Fusion, a hybrid machine learning framework that integrates chemical heuristics, engineered physical descriptors, and large language model (LLM) embeddings to accelerate the discovery of topological materials. By incorporating features such as space group symmetry, valence electron configurations, and composition-derived metrics, TXL Fusion classifies materials across trivial, TSM, and TI categories with improved accuracy and generalization compared to conventional approaches. The framework successfully identified new candidates, with representative cases further validated through density functional theory (DFT), confirming its predictive robustness. By uniting data-driven learning with chemical intuition, TXL Fusion enables rapid and interpretable exploration of complex materials spaces, establishing a scalable paradigm for the intelligent discovery of next-generation topological and quantum materials.

Topological materials--including insulators (TIs) and semimetals (TSMs)--hold immense promise for quantum technologies, yet their discovery remains constrained by the high computational cost of first-principles calculations and the slow, resource-intensive nature of experimental synthesis. Here, we introduce TXL Fusion, a hybrid machine learning framework that integrates chemical heuristics, engineered physical descriptors, and large language model (LLM) embeddings to accelerate the discovery of topological materials. By incorporating features such as space group symmetry, valence electron configurations, and composition-derived metrics, TXL Fusion classifies materials across trivial, TSM, and TI categories with improved accuracy and generalization compared to conventional approaches. The framework successfully identified new candidates, with representative cases further validated through density functional theory (DFT), confirming its predictive robustness. By uniting data-driven learning with chemical intuition, TXL Fusion enables rapid and interpretable exploration of complex materials spaces, establishing a scalable paradigm for the intelligent discovery of next-generation topological and quantum materials.

Accurate determination of crystal structures is central to materials science, underpinning the understanding of composition-structure-property relationships and the discovery of new materials. Powder X-ray diffraction is a key technique in this pursuit due to its versatility and reliability. However, current analysis pipelines still rely heavily on expert knowledge and slow iterative fitting, limiting their scalability in high-throughput and autonomous settings. Here, we introduce a physics-guided contrastive learning framework termed as XCCP. It aligns powder diffraction patterns with candidate crystal structures in a shared embedding space to enable efficient structure retrieval and symmetry recognition. The XRD encoder employs a dual-expert design with a Kolmogorov-Arnold Network projection head, one branch emphasizes low angle reflections reflecting long-range order, while the other captures dense high angle peaks shaped by symmetry. Coupled with a crystal graph encoder, contrastive pretraining yields physically grounded representations. XCCP demonstrates strong performance across tasks, with structure retrieval reaching 0.89 and space group identification attains 0.93 accuracy. The framework further generalizes to compositionally similar multi principal element alloys and demonstrates zero-shot transfer to experimental patterns. These results establish XCCP as a robust, interpretable, and scalable approach that offers a new paradigm for X-ray diffraction analysis. XCCP facilitates high-throughput screening, rapid structural validation and integration into autonomous laboratories.

The growing number of ceramics exhibiting bulk plasticity at room temperature has renewed interest in revisiting plastic deformation and dislocation-mediated mechanical and functional properties in these materials. In this work, a data-driven approach is employed to identify the key parameters governing room-temperature bulk plasticity in ceramics. The model integrates an existing dataset of 55 ceramic materials, 38 plastically deformable and 17 brittle, and achieves accurate classification of bulk plasticity. The analysis reveals several key parameters essential for predicting bulk plasticity: i) Poisson's ratio and Pugh's ratio as macroscopic indicators reflecting the balance between shear and volumetric deformation resistance, and ii) Burgers vector, crystal structure and melting temperature as crystallographic descriptors associated with lattice geometry, slip resistance and thermal stability, and iii) Bader charge as a microscopic measure of bonding character. Together, these parameters define a multiscale descriptor space linking intrinsic materials properties to bulk room-temperature plasticity in ceramics, bridging the gap between empirical ductility criteria and atomistic mechanisms of dislocation-mediated plasticity. While preliminary, this study provides the first systematic, data-driven mapping of the governing factors of ceramic plasticity. The resulting framework establishes a foundation for unifying experimental and computational studies through shared datasets and descriptors, fostering collective progress toward understanding and designing intrinsically ductile ceramics.

Negative refraction of coherent phonons is crucial for thermal management and quantum information processing, but it remains unrealized because achieving the suitable dispersion for negative refraction simultaneously with long-range coherence is challenging. In this letter, we overcome this limitation by introducing a momentum compensation mechanism mediated by discrete translational symmetry. Interfacial reciprocal lattice vectors provide momentum compensation during phonon tunneling and induce asymmetric mode matching, resulting in negative refraction without requiring strong dispersion anisotropy or a negative-curvature band. Using non-equilibrium Green's function formalism, we demonstrate coherent negative refraction of isotropic acoustic phonons in graphene/hexagonal boron nitride heterostructures. This general mechanism enables active control of phonon flow via interfacial design, paving the way for applications in atomic-scale phonon lenses and directional thermal transport.

In materials with broken inversion symmetry, nonreciprocal magneto-transport (NRMT) manifests as a bilinear dependence of charge conductivity on applied electric (E) and magnetic (B) fields. This phenomenon is deeply rooted in symmetry and electronic quantum geometry, holding promise for novel rectification and detector technologies. Existing experimental studies generally attribute NRMT to Zeeman-driven mechanisms and exhibit quadratic scaling with conductivity. Here, we report a previously unknown NRMT microscopic mechanism - Lorentz skew scattering (LSK) - revealed through the discovery of an unprecedented quartic scaling law of NRMT as well as quantitative agreement between theory and experiment in BiTeBr. LSK emerges from the interplay of Lorentz force and skew scattering, bridging classical field effect to quantum scattering effect on the Fermi surface. We demonstrate that the LSK dominates NRMT in BiTeBr, and elucidate that this dominance over other possible contributions stems from high mobility and strong Rashba splitting. The finding of LSK mechanism is of unique importance because it unveils the leading NRMT effect in high-mobility systems and suggests a universal principle towards strong NRMT by enhancing electronic relaxation time in topological materials, rendering a new designing idea for low-dissipation rectifiers and high-performance quantum electronics.

Designing materials with targeted properties remains challenging due to the vastness of chemical space and the scarcity of property-labeled data. While recent advances in generative models offer a promising way for inverse design, most approaches require large datasets and must be retrained for every new target property. Here, we introduce the EGMOF (Efficient Generation of MOFs), a hybrid diffusion-transformer framework that overcomes these limitations through a modular, descriptor-mediated workflow. EGMOF decomposes inverse design into two steps: (1) a one-dimensional diffusion model (Prop2Desc) that maps desired properties to chemically meaningful descriptors followed by (2) a transformer model (Desc2MOF) that generates structures from these descriptors. This modular hybrid design enables minimal retraining and maintains high accuracy even under small-data conditions. On a hydrogen uptake dataset, EGMOF achieved over 95% validity and 84% hit rate, representing significant improvements of up to 57% in validity and 14% in hit rate compared to existing methods, while remaining effective with only 1,000 training samples. Moreover, our model successfully performed conditional generation across 29 diverse property datasets, including CoREMOF, QMOF, and text-mined experimental datasets, whereas previous models have not. This work presents a data-efficient, generalizable approach to the inverse design of diverse MOFs and highlights the potential of modular inverse design workflows for broader materials discovery.

Vision-Language Models (VLMs) have shown strong performance in tasks like visual question answering and multimodal text generation, but their effectiveness in scientific domains such as materials science remains limited. While some machine learning methods have addressed specific challenges in this field, there is still a lack of foundation models designed for broad tasks like polymer property prediction using multimodal data. In this work, we present a multimodal polymer dataset to fine-tune VLMs through instruction-tuning pairs and assess the impact of multimodality on prediction performance. Our fine-tuned models, using LoRA, outperform unimodal and baseline approaches, demonstrating the benefits of multimodal learning. Additionally, this approach reduces the need to train separate models for different properties, lowering deployment and maintenance costs.

Second-order nonlinear optical materials enable frequency doubling of light (second-harmonic generation, SHG), which is essential for optoelectronic applications ranging from materials characterization to quantum technologies. However, comparing SHG performance across materials remains challenging as the second-order nonlinear susceptibility $\chi^{(2)}$ spans several orders of magnitude and strongly depends on the band gap $E_g$. To address this, we empirically validate a theoretical upper bound on $\chi^{(2)}$ using new databases of \textit{ab initio}-computed nonlinear optical (NLO) properties. We then formulate a normalized descriptor, $\hat{d}$, which expresses the NLO response of a material relative to the band gap-dependent physical limit. We show that $\hat{d}$ exhibits a similar distribution across a wide range of band gap energies. This universality supports the use of $\hat{d}$ as a robust, generalizable descriptor for data-driven and chemistry-informed machine learning models of NLO response, enabling accelerated materials discovery and optimization across broad application frequencies.

In the era of AI-driven science and engineering, we often want to design discrete objects in silico according to user-specified properties. For example, we may wish to design a protein to bind its target, arrange components within a circuit to minimize latency, or find materials with certain properties. Given a property predictive model, in silico design typically involves training a generative model over the design space (e.g., protein sequence space) to concentrate on designs with the desired properties. Distributional optimization -- which can be formalized as an estimation of distribution algorithm or as reinforcement learning policy optimization -- finds the generative model that maximizes an objective function in expectation. Optimizing a distribution over discrete-valued designs is in general challenging because of the combinatorial nature of the design space. However, many property predictors in scientific applications are decomposable in the sense that they can be factorized over design variables in a way that could in principle enable more effective optimization. For example, amino acids at a catalytic site of a protein may only loosely interact with amino acids of the rest of the protein to achieve maximal catalytic activity. Current distributional optimization algorithms are unable to make use of such decomposability structure. Herein, we propose and demonstrate use of a new distributional optimization algorithm, Decomposition-Aware Distributional Optimization (DADO), that can leverage any decomposability defined by a junction tree on the design variables, to make optimization more efficient. At its core, DADO employs a soft-factorized "search distribution" -- a learned generative model -- for efficient navigation of the search space, invoking graph message-passing to coordinate optimization across linked factors.

Data-driven scientific discovery requires iterative cycles of literature search, hypothesis generation, and data analysis. Substantial progress has been made towards AI agents that can automate scientific research, but all such agents remain limited in the number of actions they can take before losing coherence, thus limiting the depth of their findings. Here we present Kosmos, an AI scientist that automates data-driven discovery. Given an open-ended objective and a dataset, Kosmos runs for up to 12 hours performing cycles of parallel data analysis, literature search, and hypothesis generation before synthesizing discoveries into scientific reports. Unlike prior systems, Kosmos uses a structured world model to share information between a data analysis agent and a literature search agent. The world model enables Kosmos to coherently pursue the specified objective over 200 agent rollouts, collectively executing an average of 42,000 lines of code and reading 1,500 papers per run. Kosmos cites all statements in its reports with code or primary literature, ensuring its reasoning is traceable. Independent scientists found 79.4% of statements in Kosmos reports to be accurate, and collaborators reported that a single 20-cycle Kosmos run performed the equivalent of 6 months of their own research time on average. Furthermore, collaborators reported that the number of valuable scientific findings generated scales linearly with Kosmos cycles (tested up to 20 cycles). We highlight seven discoveries made by Kosmos that span metabolomics, materials science, neuroscience, and statistical genetics. Three discoveries independently reproduce findings from preprinted or unpublished manuscripts that were not accessed by Kosmos at runtime, while four make novel contributions to the scientific literature.

Sulfur-doped black phosphorus (b-P) has recently emerged as a promising candidate for next-generation electronic and optoelectronic technologies owing to its enhanced environmental stability and tunable electronic properties. In this work, we systematically investigate the effects of sulfur substitution on the elastic, mechanical, and electronic properties of b-P, with a particular focus on its auxetic behavior (that is, negative Poisson's ratio), using first-principles density functional theory calculations. Our results unveil the fundamental origin of the intrinsic auxetic response in pristine b-P and elucidate how sulfur incorporation alters this behavior. We find that sulfur atoms distort the characteristic bow-tie structural motif responsible for the negative Poisson's ratio in b-P, thereby suppressing the in-plane auxeticity. Moreover, the resulting charge redistribution also effectively quenches the out-of-plane auxetic response of b-P. With increasing sulfur content, the bulk modulus and Poisson's ratio increase, whereas the Young's modulus, shear modulus, and Debye temperature decrease. Additionally, sulfur substitution suppresses the semiconducting properties of b-P, giving rise to metallicity. These findings highlight that although sulfur substitution enhances the environmental stability of b-P, it also substantially modifies its elastic and mechanical properties, particularly the auxetic behavior, which is an important consideration in the design of nanoscale electronic devices.

The performance of bio-inspired fibrillar adhesives can be compromised by surface roughness, manufacturing imperfections or impurities. Previous studies investigated the cases of distributed defects on the array, and defects at the level of single fibrils. However, the influence of localized, macroscopic defects remains largely unexplored. Using numerical simulations of a discrete mechanical model for a fibrillar adhesive with a thick backing layer, we investigate how the size and location of a single circular defect affect the established scaling law between the adhesion force ($F$) and the effective compliance of the system ($β$), \ie, $F \propto β^{-1/2}$. We find that edge defects are generally more detrimental than central ones, as they act as pre-cracks that amplify stress concentrations at the adhesive's edge, accelerating a crack-like failure. Consequently, the established adhesion scaling law is preserved, with the defect only reducing the effective contact area. In contrast, a central defect fundamentally alters the mechanics of detachment. By transforming the contact geometry into an annulus, it promotes more uniform load sharing across the remaining fibrils and mitigates the edge-dominated failure mechanism. This change makes the adhesive strength less sensitive to the compliance of the system, as reflected by a less negative scaling exponent. The transition between these two regimes appears to occur for defects whose boundary merges with the one of the adhesive. These results provide practical guidance for the design, engineering and quality control of bio-inspired fibrillar adhesives.

Polymer packaging plays a crucial role in food preservation but poses major challenges in recycling and environmental persistence. To address the need for sustainable, high-performance alternatives, we employed a polymer informatics workflow to identify single- and multi-layer drop-in replacements for polymer-based packaging materials. Machine learning (ML) models, trained on carefully curated polymer datasets, predicted eight key properties across a library of approximately 7.4 million ring-opening polymerization (ROP) polymers generated by virtual forward synthesis (VFS). Candidates were prioritized by the enthalpy of polymerization, a critical metric for chemical recyclability. This screening yielded thousands of promising candidates, demonstrating the feasibility of replacing diverse packaging architectures. We then experimentally validated poly(p-dioxanone) (poly-PDO), an existing ROP polymer whose barrier performance had not been previously reported. Validation showed that poly-PDO exhibits strong water barrier performance, mechanical and thermal properties consistent with predictions, and excellent chemical recyclability (95% monomer recovery), thereby meeting the design targets and underscoring its potential for sustainable packaging. These findings highlight the power of informatics-driven approaches to accelerate the discovery of sustainable polymers by uncovering opportunities in both existing and novel chemistries.

Computational materials discovery relies on the generation of plausible crystal structures. The plausibility is typically judged through density functional theory methods which, while typically accurate at zero Kelvin, often favor low-energy structures that are not experimentally accessible. We develop a combined compositional and structural synthesizability score which provides an accurate way of predicting which compounds can actually be synthesized in a laboratory. We use it to evaluate non-synthesized structures from the Materials Project, GNoME, and Alexandria, and identified several hundred highly synthesizable candidates. We then predict synthesis pathways, conduct corresponding experiments, and characterize the products across 16 targets, successfully synthesizing 7 of 16. The entire experimental process was completed in only three days. Our results highlight omissions in lists of known synthesized structures, deliver insights into the practical utility of current materials databases, and showcase the central role synthesizability prediction can play in materials discovery.

Refractory Complex Concentrated Alloys (RCCAs) can exhibit exceptional high-temperature strength, making such alloys promising candidates for high-temperature structural applications. However, current RCCAs do not possess the high-temperature oxidation resistance required to survive in oxidizing environments for more than a few hours at or above 1000$^\circ$C, without relying primarily on an environmental barrier coating. Here, we present a machine-learning framework designed to predict the oxidation-induced specific mass changes of RCCAs exposed for 24 h at 1000$^\circ$C in air, in order to support the search for oxidation-resistant alloys over a wide range of compositions. A database was constructed of experimental specific mass change data, upon oxidation at 900-1000$^\circ$C for 24 h in air, for 77 compositions comprised of simple elements, binary alloys, and higher-order elemental systems. We then developed a Gaussian Process Regression (GPR) model with physics-informed descriptors based on oxidation products, capturing the fundamental chemistry of oxide formation and stability. Application of this GPR model to the database yielded a MAE (mean absolute error) test score of 5.78 mg/cm$^2$, which was a significant improvement in accuracy relative to models only utilizing traditional alloy-based descriptors. Our model was used to screen over 5,100 quaternary RCCAs, revealing compositions with significantly lower predicted specific mass changes compared to existing literature sources. Overall, this work establishes a versatile and efficient strategy to accelerate the discovery of next-generation RCCAs with enhanced resistance to extreme environments.

Thermally activated delayed fluorescence (TADF) emitters are essential for next-generation, high-efficiency organic light-emitting diodes (OLEDs), yet their rational design is hampered by the high computational cost of accurate excited-state predictions. Here, we present a comprehensive benchmark study validating semi-empirical extended tight-binding (xTB) methods -- specifically sTDA-xTB and sTD-DFT-xTB -- for the high-throughput screening of TADF materials. Using an unprecedentedly large dataset of \num{747} experimentally characterized emitters, our framework demonstrates a computational cost reduction of over \qty{99}{\percent} compared to conventional TD-DFT, while maintaining strong internal consistency between methods (Pearson $r \approx \num{0.82}$ for \deltaest), validating their utility for relative molecular ranking. Validation against \num{312} experimental \deltaest values reveals a mean absolute error of approximately \qty{0.17}{\electronvolt}, a discrepancy attributed to the vertical approximation inherent to the HTS protocol, underscoring the methods' role in screening rather than quantitative prediction. Through large-scale data analysis, we statistically validate key design principles, confirming the superior performance of Donor-Acceptor-Donor (D-A-D) architectures and identifying an optimal D-A torsional angle range of \qtyrange{50}{90}{\degree} for efficient TADF. Principal Component Analysis reveals that the complex property space is fundamentally low-dimensional, with three components capturing nearly \qty{90}{\percent} of the variance. This work establishes these semi-empirical methods as powerful, cost-effective tools for accelerating TADF discovery and provides a robust set of data-driven design rules and methodological guidelines for the computational materials science community.

Poly(ethylene terephthalate) (PET), a widely used thermoplastic in packaging, textiles, and engineering applications, is valued for its strength, clarity, and chemical resistance. Increasing environmental impact concerns and regulatory pressures drive the search for alternatives with comparable or superior performance. We present an AI-driven polymer design pipeline employing virtual forward synthesis (VFS) to generate PET-replacement copolymers. Inspired by the esterification route of PET synthesis, we systematically combined a down-selected set of Toxic Substances Control Act (TSCA)-listed monomers to create 12,100 PET-like polymers. Machine learning models predicted glass transition temperature (Tg), band gap, and tendency to crystallize, for all designs. Multi-objective screening identified 1,108 candidates predicted to match or exceed PET in $T_{\rm g}$ and band gap, including the ``rediscovery'' of other known commercial PET-alternate polymers (e.g., PETG, Tritan, Ecozen) that provide retrospective validation of our design pipeline, demonstrating a capability to rapidly design experimentally feasible polymers at a scale. Furthermore, selected, entirely new (previously unknown) candidates designed here have been synthesized and characterized, providing a definitive validation of the design framework.

The conjugate gradient solver (CG) is a prevalent method for solving symmetric and positive definite linear systems Ax=b, where effective preconditioners are crucial for fast convergence. Traditional preconditioners rely on prescribed algorithms to offer rigorous theoretical guarantees, while limiting their ability to exploit optimization from data. Existing learning-based methods often utilize Graph Neural Networks (GNNs) to improve the performance and speed up the construction. However, their reliance on incomplete factorization leads to significant challenges: the associated triangular solve hinders GPU parallelization in practice, and introduces long-range dependencies which are difficult for GNNs to model. To address these issues, we propose a learning-based method to generate GPU-friendly preconditioners, particularly using GNNs to construct Sparse Approximate Inverse (SPAI) preconditioners, which avoids triangular solves and requires only two matrix-vector products at each CG step. The locality of matrix-vector product is compatible with the local propagation mechanism of GNNs. The flexibility of GNNs also allows our approach to be applied in a wide range of scenarios. Furthermore, we introduce a statistics-based scale-invariant loss function. Its design matches CG's property that the convergence rate depends on the condition number, rather than the absolute scale of A, leading to improved performance of the learned preconditioner. Evaluations on three PDE-derived datasets and one synthetic dataset demonstrate that our method outperforms standard preconditioners (Diagonal, IC, and traditional SPAI) and previous learning-based preconditioners on GPUs. We reduce solution time on GPUs by 40%-53% (68%-113% faster), along with better condition numbers and superior generalization performance. Source code available at https://github.com/Adversarr/LearningSparsePreconditioner4GPU

The excitonic insulator (EI) is a more than 60-year-old theoretical proposal that yet remains elusive. It is a purely quantum phenomenon involving the spontaneous generation of excitons in quantum mechanics and the spontaneous condensation of excitons in quantum statistics. At this point, the excitons represent the ground state rather than the conventional excited state. Thus, the scarcity of candidate materials is a key factor contributing to the lack of recognized EI to date. In this review, we begin with the birth of EI, presenting the current state of the field and the main challenges it faces. We then focus on recent advances in the discovery and design of EIs based on the first-principles Bethe-Salpeter scheme, in particular the dark-exciton rule guided screening of materials. It not only opens up new avenues for realizing excitonic instability in direct-gap and wide-gap semiconductors, but also leads to the discovery of novel quantum states of matter such as half-EIs and spin-triplet EIs. Finally, we will look ahead to possible research pathways leading to the first recognized EI, both computationally and theoretically.

Intensive studies have been made on antiferromagnets as candidate materials for next generation memory bits due to their ultrafast dynamics reaching picosecond time scales. Recent demonstrations of electrical bidirectional switching of antiferromagnetic states have attracted significant attention. However, under the presence of significant Joule heating that destabilizes the magnetic order, the timescales associated with the switching can be limited to nanoseconds or longer. Here, we present the observation of a crossover in the switching behavior of the chiral antiferromagnet Mn3Sn by tuning the magnetic layer thickness. While Joule heating interferes with switching in thicker devices, we find clear signatures of an intrinsic spin-orbit torque mechanism as the thickness is reduced, avoiding the heating effect. The suppression of heating enables switching without significant attenuation of the readout signal using pulses shorter than those required by temperature-assisted mechanisms. The crossover into the spin-orbit torque switching behavior clarifies the potential for achieving ultrafast switching as expected from the picosecond spin dynamics of antiferromagnets. Our results lay the groundwork for designing antiferromagnetic memory devices that can operate at ultrafast timescales.

We present Denario, an AI multi-agent system designed to serve as a scientific research assistant. Denario can perform many different tasks, such as generating ideas, checking the literature, developing research plans, writing and executing code, making plots, and drafting and reviewing a scientific paper. The system has a modular architecture, allowing it to handle specific tasks, such as generating an idea, or carrying out end-to-end scientific analysis using Cmbagent as a deep-research backend. In this work, we describe in detail Denario and its modules, and illustrate its capabilities by presenting multiple AI-generated papers generated by it in many different scientific disciplines such as astrophysics, biology, biophysics, biomedical informatics, chemistry, material science, mathematical physics, medicine, neuroscience and planetary science. Denario also excels at combining ideas from different disciplines, and we illustrate this by showing a paper that applies methods from quantum physics and machine learning to astrophysical data. We report the evaluations performed on these papers by domain experts, who provided both numerical scores and review-like feedback. We then highlight the strengths, weaknesses, and limitations of the current system. Finally, we discuss the ethical implications of AI-driven research and reflect on how such technology relates to the philosophy of science. We publicly release the code at https://github.com/AstroPilot-AI/Denario. A Denario demo can also be run directly on the web at https://huggingface.co/spaces/astropilot-ai/Denario, and the full app will be deployed on the cloud.

Modern materials science generates vast and diverse datasets from both experiments and computations, yet these multi-source, heterogeneous data often remain disconnected in isolated "silos". Here, we introduce MaterialsGalaxy, a comprehensive platform that deeply fuses experimental and theoretical data in condensed matter physics. Its core innovation is a structure similarity-driven data fusion mechanism that quantitatively links cross-modal records - spanning diffraction, crystal growth, computations, and literature - based on their underlying atomic structures. The platform integrates artificial intelligence (AI) tools, including large language models (LLMs) for knowledge extraction, generative models for crystal structure prediction, and machine learning property predictors, to enhance data interpretation and accelerate materials discovery. We demonstrate that MaterialsGalaxy effectively integrates these disparate data sources, uncovering hidden correlations and guiding the design of novel materials. By bridging the long-standing gap between experiment and theory, MaterialsGalaxy provides a new paradigm for data-driven materials research and accelerates the discovery of advanced materials.

In the past decade, moir\'e materials have revolutionized how we engineer and control quantum phases of matter. Among incommensurate materials, moir\'e materials are aperiodic composite crystals whose long-wavelength moir\'e superlattices enable tunable properties without chemically modifying their layers. To date, nearly all reports of moir\'e materials have investigated van der Waals heterostructures assembled far from thermodynamic equilibrium. Here we introduce a conceptually new approach to synthesizing high-mobility moir\'e materials in thermodynamic equilibrium. We report a new family of foliated superlattice materials (Sr$_6$TaS$_8$)$_{1+\delta}$(TaS$_2$)$_8$ that are exfoliatable van der Waals crystals with atomically incommensurate lattices. Lattice mismatches between alternating layers generate moir\'e superlattices, analogous to those of 2D moir\'e heterobilayers, that are coherent throughout these crystals and are tunable through their synthesis conditions without altering their chemical composition. High-field quantum oscillation measurements map the complex Fermiology of these moir\'e metals, which can be tuned via the moir\'e superlattice structure. We find that the Fermi surface of the structurally simplest moir\'e metal is comprised of over 40 distinct cross-sectional areas, the most observed in any material to our knowledge. This can be naturally understood by postulating that bulk moir\'e materials can encode electronic properties of higher-dimensional superspace crystals in ways that parallel well-established crystallographic methods used for incommensurate lattices. More broadly, our work demonstrates a scalable synthesis approach potentially capable of producing moir\'e materials for electronics applications and evidences a novel material design concept for accessing a broad range of physical phenomena proposed in higher dimensions.

The interaction between hydrogen atoms and carbon nanoparticles is a fundamental process governing the properties of carbonaceous materials in environments ranging from combustion systems to the interstellar medium. This study investigates the effects of controlled atomic hydrogen exposure on young and mature soot nanoparticles, generated in premixed ethylene-air flames, and deposited on substrates. We employed a multi-technique approach to characterize the chemical, mechanical, and electrical evolution of the films. In-situ infrared spectroscopy revealed non-monotonic behavior: an initial increase in aliphatic CH bonds was observed, followed by a decrease at higher hydrogen fluences. This was accompanied by a continuous decrease in the aromatic C=C signal. Atomic force microscopy showed a significant increase in the Young's modulus of the film for both sample types after hydrogenation. This mechanical change was correlated with an increase in the I(D)/I(G) ratio from Raman spectroscopy. Furthermore, both macroscopic current vs. voltage and local scanning tunneling spectroscopy measurements demonstrated a notable increase in electrical conductivity. For single just-formed soot particles, moreover, a hydrogen-induced transformation from a semiconductive to a semi-metallic nature was observed. The collective evidence points towards an H-induced CC cross-linking mechanism within the nanoparticle films. We propose that atomic hydrogen facilitates the formation of radical sites, which promotes covalent bond formation between adjacent particles or molecular units, creating a more interconnected and rigid network, with smaller interlayer distance. These findings provide crucial insights into the structural evolution of carbonaceous materials in hydrogen-rich environments, with direct implications for understanding soot formation and for the tailored design of carbon-based materials.

Large language models are now integrated into many scientific workflows, accelerating data analysis, hypothesis generation, and design space exploration. In parallel with this growth, there is a growing need to carefully evaluate whether models accurately capture domain-specific knowledge and notation, since general-purpose benchmarks rarely reflect these requirements. This gap is especially clear in quantum science, which features non-intuitive phenomena and requires advanced mathematics. In this study, we introduce QuantumBench, a benchmark for the quantum domain that systematically examine how well LLMs understand and can be applied to this non-intuitive field. Using publicly available materials, we compiled approximately 800 questions with their answers spanning nine areas related to quantum science and organized them into an eight-option multiple-choice dataset. With this benchmark, we evaluate several existing LLMs and analyze their performance in the quantum domain, including sensitivity to changes in question format. QuantumBench is the first LLM evaluation dataset built for the quantum domain, and it is intended to guide the effective use of LLMs in quantum research.

This paper systematically reviews the research progress and application prospects of machine learning technologies in the field of polymer materials. Currently, machine learning methods are developing rapidly in polymer material research; although they have significantly accelerated material prediction and design, their complexity has also caused difficulties in understanding and application for researchers in traditional fields. In response to the above issues, this paper first analyzes the inherent challenges in the research and development of polymer materials, including structural complexity and the limitations of traditional trial-and-error methods. To address these problems, it focuses on introducing key basic technologies such as molecular descriptors and feature representation, data standardization and cleaning, and records a number of high-quality polymer databases. Subsequently, it elaborates on the key role of machine learning in polymer property prediction and material design, covering the specific applications of algorithms such as traditional machine learning, deep learning, and transfer learning; further, it deeply expounds on data-driven design strategies, such as reverse design, high-throughput virtual screening, and multi-objective optimization. The paper also systematically introduces the complete process of constructing high-reliability machine learning models and summarizes effective experimental verification, model evaluation, and optimization methods. Finally, it summarizes the current technical challenges in research, such as data quality and model generalization ability, and looks forward to future development trends including multi-scale modeling, physics-informed machine learning, standardized data sharing, and interpretable machine learning.

Heat conduction and radiation are two of the three fundamental modes of heat transfer, playing a critical role in a wide range of scientific and engineering applications ranging from energy systems to materials science. However, traditional physics-based simulation methods for modeling these processes often suffer from prohibitive computational costs. In recent years, the rapid advancements in Artificial Intelligence (AI) and machine learning (ML) have demonstrated remarkable potential in the modeling of nanoscale heat conduction and radiation. This review presents a comprehensive overview of recent AI-driven developments in modeling heat conduction and radiation at the nanoscale. We first discuss the ML techniques for predicting phonon properties, including phonon dispersion and scattering rates, which are foundational for determining material thermal properties. Next, we explore the role of machine-learning interatomic potentials (MLIPs) in molecular dynamics simulations and their applications to bulk materials, low-dimensional systems, and interfacial transport. We then review the ML approaches for solving radiative heat transfer problems, focusing on data-driven solutions to Maxwell's equations and the radiative transfer equation. We further discuss the ML-accelerated inverse design of radiative energy devices, including optimization-based and generative model-based methods. Finally, we discuss open challenges and future directions, including data availability, model generalization, uncertainty quantification, and interpretability. Through this survey, we aim to provide a foundational understanding of how AI techniques are reshaping thermal science and guiding future research in nanoscale heat transfer.

A triangular Co-ion lattice intercalated between 1-H NbS$_2$ layers can exhibit a large anomalous Hall effect (AHE) due to the finite scalar spin chirality originating from the non-coplanar $3q$ ordering of Co spins. This large AHE occurs when the scalar spin chirality is uniform in all Co layers, as indeed found in the Co$_{1/3}$NbS$_2$ case [Phys. Rev. Mater. 6, 024201 (2022)]. However, if the spin chirality were staggered with the opposite signs in the adjacent Co layers, the net AHE would disappear, yielding instead the topological magneto-electric effect. Here, we theoretically verify that a transverse electric field generates a finite orbital magnetization under such conditions, consistent with the axion-like coupling. Using first-principles calculations, we show that the resulting magneto-electric coupling, $α^{zz}$ can be as large as 0.9 $e^2/2h$. We also demonstrate that the inter-layer magnetic coupling in these materials can be tuned by strain, enabling the switching between the AHE and the axionic states.

Differentiating through constrained optimization problems is increasingly central to learning, control, and large-scale decision-making systems, yet practical integration remains challenging due to solver specialization and interface mismatches. This paper presents a general and streamlined framework-an updated DiffOpt.jl-that unifies modeling and differentiation within the Julia optimization stack. The framework computes forward - and reverse-mode solution and objective sensitivities for smooth, potentially nonconvex programs by differentiating the KKT system under standard regularity assumptions. A first-class, JuMP-native parameter-centric API allows users to declare named parameters and obtain derivatives directly with respect to them - even when a parameter appears in multiple constraints and objectives - eliminating brittle bookkeeping from coefficient-level interfaces. We illustrate these capabilities on convex and nonconvex models, including economic dispatch, mean-variance portfolio selection with conic risk constraints, and nonlinear robot inverse kinematics. Two companion studies further demonstrate impact at scale: gradient-based iterative methods for strategic bidding in energy markets and Sobolev-style training of end-to-end optimization proxies using solver-accurate sensitivities. Together, these results demonstrate that differentiable optimization can be deployed as a routine tool for experimentation, learning, calibration, and design-without deviating from standard JuMP modeling practices and while retaining access to a broad ecosystem of solvers.

This paper investigates the performance of Large Language Models (LLMs) as generative optimizers for solving constrained multi-objective regression tasks, specifically within the challenging domain of inverse design (property-to-structure mapping). This problem, critical to materials informatics, demands finding complex, feasible input vectors that lie on the Pareto optimal front. While LLMs have demonstrated universal effectiveness across generative and reasoning tasks, their utility in constrained, continuous, high-dimensional numerical spaces tasks they weren't explicitly architected for remains an open research question. We conducted a rigorous comparative study between established Bayesian Optimization (BO) frameworks and a suite of fine-tuned LLMs and BERT models. For BO, we benchmarked the foundational BoTorch Ax implementation against the state-of-the-art q-Expected Hypervolume Improvement (qEHVI, BoTorchM). The generative approach involved fine-tuning models via Parameter-Efficient Fine-Tuning (PEFT), framing the challenge as a regression problem with a custom output head. Our results show that BoTorch qEHVI achieved perfect convergence (GD=0.0), setting the performance ceiling. Crucially, the best-performing LLM (WizardMath-7B) achieved a Generational Distance (GD) of 1.21, significantly outperforming the traditional BoTorch Ax baseline (GD=15.03). We conclude that specialized BO frameworks remain the performance leader for guaranteed convergence, but fine-tuned LLMs are validated as a promising, computationally fast alternative, contributing essential comparative metrics to the field of AI-driven optimization. The findings have direct industrial applications in optimizing formulation design for resins, polymers, and paints, where multi-objective trade-offs between mechanical, rheological, and chemical properties are critical to innovation and production efficiency.

Normalizing flows are deep generative models that enable efficient likelihood estimation and sampling through invertible transformations. A key challenge is to design linear layers that enhance expressiveness while maintaining efficient computation of the Jacobian determinant and inverse. We introduce a novel invertible linear layer based on the product of circulant and diagonal matrices. This decomposition reduces parameter complexity from $\mathcal{O}(n^2)$ to $\mathcal{O}(mn)$ using $m$ diagonal matrices and $m-1$ circulant matrices while still approximating general linear transformations. By leveraging the Fast Fourier Transform, our approach reduces the time complexity of matrix inversion from $\mathcal{O}(n^3)$ to $\mathcal{O}(mn\log n)$ and that of computing the log-determinant from $\mathcal{O}(n^3)$ to $\mathcal{O}(mn)$, where $n$ is the input dimension. We build upon this layer to develop Circulant-Diagonal Flow (CDFlow), which achieves strong density estimation on natural image datasets and effectively models data with inherent periodic structure. Furthermore, CDFlow significantly accelerates key operations in normalizing flows, providing practical benefits for scalable generative modeling.

Normalizing flows are deep generative models that enable efficient likelihood estimation and sampling through invertible transformations. A key challenge is to design linear layers that enhance expressiveness while maintaining efficient computation of the Jacobian determinant and inverse. We introduce a novel invertible linear layer based on the product of circulant and diagonal matrices. This decomposition reduces parameter complexity from $\mathcal{O}(n^2)$ to $\mathcal{O}(mn)$ using $m$ diagonal matrices and $m-1$ circulant matrices while still approximating general linear transformations. By leveraging the Fast Fourier Transform, our approach reduces the time complexity of matrix inversion from $\mathcal{O}(n^3)$ to $\mathcal{O}(mn\log n)$ and that of computing the log-determinant from $\mathcal{O}(n^3)$ to $\mathcal{O}(mn)$, where $n$ is the input dimension. We build upon this layer to develop Circulant-Diagonal Flow (CDFlow), which achieves strong density estimation on natural image datasets and effectively models data with inherent periodic structure. Furthermore, CDFlow significantly accelerates key operations in normalizing flows, providing practical benefits for scalable generative modeling.

We report the discovery of a geometric pathway for tuning ferroelectric properties through thermally driven reconfiguration between coexisting polar states in Li-substituted NaNbO3. Using first-principles density functional theory calculation and 7Li solid-state nuclear magnetic resonance spectroscopy measurement, we reveal that Li substitution creates two distinct polar configurations whose transformation under annealing enhances the Curie temperature and induces piezoelectric hardening. Our findings establish a geometrically-driven polar state reconfiguration mechanism, providing a general design principle for ferroics whereby macroscopic functional properties can be engineered via lattice geometry.

When two-dimensional atomic layers of different materials are brought into close proximity to form van der Waals (vdW) heterostructures, interactions between adjacent layers significantly influence their physicochemical properties. These effects seem particularly pronounced when the interface exhibits local order and near-perfect structural alignment, leading to the emergence of Moiré patterns. Using quantum mechanical density functional theory calculations, we propose a prototypical bilayer heterostructure composed of hexagonal boron nitride (hBN) and silicon carbide (SiC), characterized by a lattice mismatch of 18.77\% between their primitive unit cells. We find that the removal of boron atoms from specific lattice sites can convert the interlayer interaction from weak vdW coupling to robust localized silicon-nitrogen covalent bonding. Motivated by this, we study the binding of transition-metal adatoms and formulate design guidelines to enhance surface reactivity, thereby enabling the controlled isolation of single-metal atoms. Our machine-learning-assisted molecular dynamics simulations confirm both dynamical stability and metal anchoring feasibility at finite temperatures. Our results suggest the hBN/SiC heterostructure as a versatile platform for atomically precise transition-metal functionalization, having potential for next-generation catalytic energy-conversion technologies.

High-energy materials (HEMs) are critical for propulsion and defense domains, yet their discovery remains constrained by experimental data and restricted access to testing facilities. This work presents a novel approach toward high-energy molecules by combining Long Short-Term Memory (LSTM) networks for molecular generation and Attentive Graph Neural Networks (GNN) for property predictions. We propose a transformative embedding space construction strategy that integrates fixed SHA-256 embeddings with partially trainable representations. Unlike conventional regularization techniques, this changes the representational basis itself, reshaping the molecular input space before learning begins. Without recourse to pretraining, the generator achieves 67.5% validity and 37.5% novelty. The generated library exhibits a mean Tanimoto coefficient of 0.214 relative to training set signifying the ability of framework to generate a diverse chemical space. We identified 37 new super explosives higher than 9 km/s predicted detonation velocity.

Cesium based halide perovskites, such as CsPbI3 and CsSnI3, have emerged as exceptional candidates for next generation photovoltaic and optoelectronic technologies, but their practical application is limited by temperature dependent phase transitions and structural instabilities. Here, we develop machine learning interatomic potentials within the LATTE framework to simulate these materials with near experimental accuracy at a fraction of the computational cost compared to previous computational studies. Our molecular dynamics simulations based on the trained MLIPs reproduce energies and forces across multiple phases, enabling large scale simulations that capture cubic tetragonal orthorhombic transitions, lattice parameters, and octahedral tilting with unprecedented resolution. We find that Pb based perovskites exhibit larger octahedral tilts and higher phase transition temperatures than Sn based analogues, reflecting stronger bonding and enhanced structural stability, whereas Sn based perovskites display reduced tilts and lower barriers, suggesting tunability through compositional or interface engineering. Beyond these systems, our work demonstrates that MLIPs can bridge first principles accuracy with simulation efficiency, providing a robust framework for exploring phase stability, anharmonicity, and rational design in next generation halide perovskites.

Topolectrical circuits provide a versatile platform for exploring and simulating modern physical models. However, existing approaches suffer from incomplete programmability and ineffective feature prediction and control mechanisms, hindering the investigation of physical phenomena on an integrated platform and limiting their translation into practical applications. Here, we present a deep learning empowered programmable topolectrical circuits (DLPTCs) platform for physical modeling and analysis. By integrating fully independent, continuous tuning of both on site and off site terms of the lattice Hamiltonian, physics graph informed inverse state design, and immediate hardware verification, our system bridges the gap between theoretical modeling and practical realization. Through flexible control and adiabatic path engineering, we experimentally observe the boundary states without global symmetry in higher order topological systems, their adiabatic phase transitions, and the flat band like characteristic corresponding to Landau levels in the circuit. Incorporating a physics graph informed mechanism with a generative AI model for physics exploration, we realize arbitrary, position controllable on board Anderson localization, surpassing conventional random localization. Utilizing this unique capability with high fidelity hardware implementation, we further demonstrate a compelling cryptographic application: hash based probabilistic information encryption by leveraging Anderson localization with extensive disorder configurations, enabling secure delivery of full ASCII messages.

In this work, we demonstrate the use of CQDs synthesized from lignin - currently one of the most abundant and underutilized by-products of paper and pulp production - for advanced pH monitoring applications. The presented approach integrates green chemistry principles with an operator-friendly, low-cost, and practical solution for spatial and temporal pH measurement. CQDs functionalized with m-aminophenylboronic acid enable highly sensitive and reversible pH readouts through two complementary mechanisms: ratiometric monitoring of emission band intensities, and direct visual observation of colorimetric changes reflected in the CIE1931 chromaticity coordinates. The system achieves maximal sensitivities of 137 percent per pH unit and 49.5 percent per pH unit, respectively, while simultaneously maintaining high measurement resolution and full reproducibility of the readouts, placing it among the most effective CQD-based pH sensors reported to date. Here, we demonstrate the capability of 2D luminescent imaging of pH distributions, allowing for both spatially resolved and time-resolved monitoring. Employing just an excitation source, a digital camera or smartphone, and RGB channel analysis, the setup eliminates the necessity for specialized filters or sophisticated instrumentation. The combination of multimodal readout strategies with the capacity for large-area visualization establishes lignin-derived CQDs as a sustainable and practical platform for pH sensing. By simultaneously addressing the challenges of waste valorization and the demand for innovative sensing technologies, this solution fulfills the requirements of both environmentally responsible material design and next-generation pH sensor development.

The intrinsic coupling among electrical conductivity ($σ$), Seebeck coefficient ($S$), and lattice thermal conductivity ($κ_{\mathrm{L}}$) imposes a fundamental limit on the dimensionless figure of merit $ZT$ in thermoelectric (TE) materials. Increasing band degeneracy can effectively balance $σ$ and $S$, enabling a high power factor (PF, $S^{2}σ$). However, compounds with intrinsically large band degeneracy are scarce. Here, we present an unconventional strategy to realize elevated band degeneracy in zigzag-chain Na$_2$Au$X$ ($X$ = P, As, Sb, Bi) compounds by harnessing strong intra- and interchain orbital coupling. Pronounced hybridization between Au-$d_{z^{2}}$ and $X$-$p_{z}$ orbitals along the Au--$X$ zigzag chains, together with unexpectedly strong interchain $X$-$p_{x}/p_{y}$ coupling, produces a highly dispersive, multivalley valence band structure that supports an exceptional PF. Concurrently, the intrinsically weak interchain interactions arising from the quasi-one-dimensional framework, together with the weakened Au--$X$ and Au--Au bonds within the chains due to filling of $p$-$d^{*}$ antibonding states, result in an ultralow $κ_{\mathrm{L}}$. First-principles calculations combined with Boltzmann transport theory predict that $p$-type Na$_2$AuBi achieves a PF of $63.9\,μ\mathrm{W}\,\mathrm{cm}^{-1}\,\mathrm{K}^{-2}$, an ultralow $κ_{\mathrm{L}}$ of $0.49\,\mathrm{W}\,\mathrm{m}^{-1}\,\mathrm{K}^{-1}$, and a maximum $ZT$ of $4.7$ along the zigzag-chain direction at $800\,\mathrm{K}$. This work establishes a new design paradigm for high-efficiency TE materials by exploiting substantial orbital overlap in structurally weakly bonded, quasi-one-dimensional systems, opening promising avenues for the discovery and engineering of next-generation high-performance TE materials.

We developed DeFecT-FF, a framework for predicting the energies and ground-state configurations of native point defects, extrinsic dopants, impurities, and defect complexes in zincblende-phase Cd/Zn-Te/Se/S compounds relevant to CdTe-based solar cells. The framework combines high-throughput DFT data with crystal graph-based machine learning force fields (MLFFs) trained to reproduce DFT energies and forces. Alloying at Cd or Te sites offers a route to tune the electronic and defect properties of CdTe absorbers for improved solar efficiency. Given the vast number of possible defect types, charge states, and symmetry-breaking configurations, traditional DFT approaches are computationally prohibitive. Our dataset includes GGA-PBE and HSE06-optimized structures for bulk, alloyed, interface, and grain-boundary systems. Using active learning, we expanded the dataset and trained MLFFs to accurately predict energies across charge states. The model enabled rapid screening and discovery of new low-energy defect configurations, validated through HSE06 calculations with spin-orbit coupling. The DeFecT-FF framework is publicly available as a nanoHUB tool, allowing users to upload crystallographic files, automatically generate defects, and compute defect formation energies versus Fermi level and chemical potentials, greatly reducing the need for expensive DFT simulations.

The increasing complexity and volume of data generated by high-throughput computational materials science require robust tools to ensure their accessibility, reproducibility, and reuse. In particular, integrating the FAIR Guiding Principles (Findable, Accessible, Interoperable, and Reusable) into computational workflows is essential to enable open science practices. TribChem is an open source Python software developed for the automated simulation of solid-solid interfaces using density functional theory (DFT). While TribChem already incorporates several FAIR-aligned features, we present here a dedicated FAIR utility designed to transform TribChem results into FAIR-compliant datasets. This utility comprises two tools: fair_data.py, which automatically generates standardized machine- and human-readable outputs from the TribChem database, and retrieve_data.py, which facilitates efficient data extraction through a keyword-based interface. In this paper we show the capabilities of the fair utility with examples for bulk, surface, and interface systems. The implementation allows seamless integration with public repositories such as Zenodo, paving the way for reproducible research and fostering data-driven materials discovery.

Designing frictional interfaces to exhibit prescribed macroscopic behavior is a challenging inverse problem, made difficult by the non-uniqueness of solutions and the computational cost of contact simulations. Traditional approaches rely on heuristic search over low-dimensional parameterizations, which limits their applicability to more complex or nonlinear friction laws. We introduce a generative modeling framework using Variational Autoencoders (VAEs) to infer surface topographies from target friction laws. Trained on a synthetic dataset composed of 200 million samples constructed from a parameterized contact mechanics model, the proposed method enables efficient, simulation-free generation of candidate topographies. We examine the potential and limitations of generative modeling for this inverse design task, focusing on balancing accuracy, throughput, and diversity in the generated solutions. Our results highlight trade-offs and outline practical considerations when balancing these objectives. This approach paves the way for near-real-time control of frictional behavior through tailored surface topographies.

Amorphous solid-state electrolytes (SSEs) offer unique advantages for next-generation batteries, but their rational design is hindered by an unclear structure-property relationship. This study establishes universal design principles through atomistic simulations of 32 amorphous Li-M-X systems (M = B, Al, Si, P; X = F, Cl, Br, I, O, S, Se, N). We identify four structure types governed by a rule that saturated M-X groups with more negative charges preferentially form M-X-M chains, identify paddle-wheel and cooperative migration as two favorable transport mechanisms that are significantly enhanced in amorphous structures. We also pinpoint Oxides and fluorides as optimal for electrochemical and hydrolytic stability, and reveal bulk modulus as a simple predictor for $Li^+$ mobility. These insights are integrated into a practical design diagram, providing a novel and valuable framework for advancing high-performance amorphous SSEs.

Discovering materials with previously unreported crystal frameworks is key to achieving transformative functionality. Generative artificial intelligence offers a scalable means to propose candidate crystal structures, however existing approaches mainly reproduce decorated variants of established motifs rather than uncover new configurations. Here we develop a physics-informed diffusion method, supported by chemically grounded validation protocol, which embeds descriptors of compactness and local environment diversity to balance physical plausibility with structural novelty. Conditioning on these metrics improves generative performance across architectures, increasing the fraction of structures outside 100 most common prototypes up to 67%. When crystal structure prediction (CSP) is seeded with generative structures, most candidates (97%) are reconstructed by CSP, yielding 145 (66%) low-energy frameworks not matching any known prototypes. These results show that while generative models are not substitutes for CSP, their chemically informed, diversity-guided outputs can enhance CSP efficiency, establishing a practical generative-CSP synergy for discovery-oriented exploration of chemical space.

Recent advances in generative modeling have shown significant promise in designing novel periodic crystal structures. Existing approaches typically rely on either large language models (LLMs) or equivariant denoising models, each with complementary strengths: LLMs excel at handling discrete atomic types but often struggle with continuous features such as atomic positions and lattice parameters, while denoising models are effective at modeling continuous variables but encounter difficulties in generating accurate atomic compositions. To bridge this gap, we propose CrysLLMGen, a hybrid framework that integrates an LLM with a diffusion model to leverage their complementary strengths for crystal material generation. During sampling, CrysLLMGen first employs a fine-tuned LLM to produce an intermediate representation of atom types, atomic coordinates, and lattice structure. While retaining the predicted atom types, it passes the atomic coordinates and lattice structure to a pre-trained equivariant diffusion model for refinement. Our framework outperforms state-of-the-art generative models across several benchmark tasks and datasets. Specifically, CrysLLMGen not only achieves a balanced performance in terms of structural and compositional validity but also generates more stable and novel materials compared to LLM-based and denoisingbased models Furthermore, CrysLLMGen exhibits strong conditional generation capabilities, effectively producing materials that satisfy user-defined constraints. Code is available at https://github.com/kdmsit/crysllmgen

The energy landscape dictates pathways and outcomes in supramolecular selfassembly, yet harnessing it from the nano to the macro scales remains a major challenge. Here, we demonstrate liquid liquid phase separation (LLPS) as a powerful tool to navigate and engineer the energy landscapes of coassembly systems comprising disordered proteins and peptides. We quantitatively map the energy barriers and transition states governing structural transitions, enabling predictive on off control of assembly and hierarchical order from nano to macro scales. By integrating supramolecular biofabrication, we achieve spatially organized architectures with life like non equilibrium behaviour. Crucially, assembly stability and scalable selfsorting are shown to depend on accessing minimum energy states, regardless of whether the co assembled structures are disordered or ordered. This work establishes energy landscape mediation via LLPS as a general framework for designing lifelike, hierarchically structured materials.

In wireless communication systems, efficient and adaptive resource allocation plays a crucial role in enhancing overall Quality of Service (QoS). While centralized Multi-Agent Reinforcement Learning (MARL) frameworks rely on a central coordinator for policy training and resource scheduling, they suffer from scalability issues and privacy risks. In contrast, the Distributed Training with Decentralized Execution (DTDE) paradigm enables distributed learning and decision-making, but it struggles with non-stationarity and limited inter-agent cooperation, which can severely degrade system performance. To overcome these challenges, we propose the Multi-Agent Conditional Diffusion Model Planner (MA-CDMP) for decentralized communication resource management. Built upon the Model-Based Reinforcement Learning (MBRL) paradigm, MA-CDMP employs Diffusion Models (DMs) to capture environment dynamics and plan future trajectories, while an inverse dynamics model guides action generation, thereby alleviating the sample inefficiency and slow convergence of conventional DTDE methods. Moreover, to approximate large-scale agent interactions, a Mean-Field (MF) mechanism is introduced as an assistance to the classifier in DMs. This design mitigates inter-agent non-stationarity and enhances cooperation with minimal communication overhead in distributed settings. We further theoretically establish an upper bound on the distributional approximation error introduced by the MF-based diffusion generation, guaranteeing convergence stability and reliable modeling of multi-agent stochastic dynamics. Extensive experiments demonstrate that MA-CDMP consistently outperforms existing MARL baselines in terms of average reward and QoS metrics, showcasing its scalability and practicality for real-world wireless network optimization.

In wireless communication systems, efficient and adaptive resource allocation plays a crucial role in enhancing overall Quality of Service (QoS). Compared to the conventional Model-Free Reinforcement Learning (MFRL) scheme, Model-Based RL (MBRL) first learns a generative world model for subsequent planning. The reuse of historical experience in MBRL promises more stable training behavior, yet its deployment in large-scale wireless networks remains challenging due to high-dimensional stochastic dynamics, strong inter-agent cooperation, and communication constraints. To overcome these challenges, we propose the Multi-Agent Conditional Diffusion Model Planner (MA-CDMP) for decentralized communication resource management. Built upon the Distributed Training with Decentralized Execution (DTDE) paradigm, MA-CDMP models each communication node as an autonomous agent and employs Diffusion Models (DMs) to capture and predict environment dynamics. Meanwhile, an inverse dynamics model guides action generation, thereby enhancing sample efficiency and policy scalability. Moreover, to approximate large-scale agent interactions, a Mean-Field (MF) mechanism is introduced as an assistance to the classifier in DMs. This design mitigates inter-agent non-stationarity and enhances cooperation with minimal communication overhead in distributed settings. We further theoretically establish an upper bound on the distributional approximation error introduced by the MF-based diffusion generation, guaranteeing convergence stability and reliable modeling of multi-agent stochastic dynamics. Extensive experiments demonstrate that MA-CDMP consistently outperforms existing MARL baselines in terms of average reward and QoS metrics, showcasing its scalability and practicality for real-world wireless network optimization.

Structural topology optimization (TO) is central to engineering design but remains computationally intensive due to complex physics and hard constraints. Existing deep-learning methods are limited to fixed square grids, a few hand-coded boundary conditions, and post-hoc optimization, preventing general deployment. We introduce Optimize Any Topology (OAT), a foundation-model framework that directly predicts minimum-compliance layouts for arbitrary aspect ratios, resolutions, volume fractions, loads, and fixtures. OAT combines a resolution- and shape-agnostic autoencoder with an implicit neural-field decoder and a conditional latent-diffusion model trained on OpenTO, a new corpus of 2.2 million optimized structures covering 2 million unique boundary-condition configurations. On four public benchmarks and two challenging unseen tests, OAT lowers mean compliance up to 90% relative to the best prior models and delivers sub-1 second inference on a single GPU across resolutions from 64 x 64 to 256 x 256 and aspect ratios as high as 10:1. These results establish OAT as a general, fast, and resolution-free framework for physics-aware topology optimization and provide a large-scale dataset to spur further research in generative modeling for inverse design. Code & data can be found at https://github.com/ahnobari/OptimizeAnyTopology.

Materials discovery requires navigating vast chemical and structural spaces while satisfying multiple, often conflicting, objectives. We present LLM-guided Evolution for MAterials discovery (LLEMA), a unified framework that couples the scientific knowledge embedded in large language models with chemistry-informed evolutionary rules and memory-based refinement. At each iteration, an LLM proposes crystallographically specified candidates under explicit property constraints; a surrogate-augmented oracle estimates physicochemical properties; and a multi-objective scorer updates success/failure memories to guide subsequent generations. Evaluated on 14 realistic tasks that span electronics, energy, coatings, optics, and aerospace, LLEMA discovers candidates that are chemically plausible, thermodynamically stable, and property-aligned, achieving higher hit rates and improved Pareto front quality relative to generative and LLM-only baselines. Ablation studies confirm the importance of rule-guided generation, memory-based refinement, and surrogate prediction. By enforcing synthesizability and multi-objective trade-offs, LLEMA provides a principled approach to accelerating practical materials discovery. Project website: https://scientific-discovery.github.io/llema-project/

Materials discovery requires navigating vast chemical and structural spaces while satisfying multiple, often conflicting, objectives. We present LLM-guided Evolution for MAterials design (LLEMA), a unified framework that couples the scientific knowledge embedded in large language models with chemistry-informed evolutionary rules and memory-based refinement. At each iteration, an LLM proposes crystallographically specified candidates under explicit property constraints; a surrogate-augmented oracle estimates physicochemical properties; and a multi-objective scorer updates success/failure memories to guide subsequent generations. Evaluated on 14 realistic tasks spanning electronics, energy, coatings, optics, and aerospace, LLEMA discovers candidates that are chemically plausible, thermodynamically stable, and property-aligned, achieving higher hit-rates and stronger Pareto fronts than generative and LLM-only baselines. Ablation studies confirm the importance of rule-guided generation, memory-based refinement, and surrogate prediction. By enforcing synthesizability and multi-objective trade-offs, LLEMA delivers a principled pathway to accelerate practical materials discovery. Code: https://github.com/scientific-discovery/LLEMA

Quantum imaginary time evolution (QITE) algorithm is one of the most promising variational quantum algorithms (VQAs), bridging the current era of Noisy Intermediate-Scale Quantum devices and the future of fully fault-tolerant quantum computing. Although practical demonstrations of QITE and its potential advantages over the general VQA trained with vanilla gradient descent (GD) in certain tasks have been reported, a first-principle, theoretical understanding of QITE remains limited. Here, we aim to develop an analytic theory for the dynamics of QITE. First, we show that QITE can be interpreted as a form of a general VQA trained with Quantum Natural Gradient Descent (QNGD), where the inverse quantum Fisher information matrix serves as the learning-rate tensor. This equivalence is established not only at the level of gradient update rules, but also through the action principle: the variational principle can be directly connected to the geometric geodesic distance in the quantum Fisher information metric, up to an integration constant. Second, for wide quantum neural networks, we employ the quantum neural tangent kernel framework to construct an analytic model for QITE. We prove that QITE always converges faster than GD-based VQA, though this advantage is suppressed by the exponential growth of Hilbert space dimension. This helps explain certain experimental results in quantum computational chemistry. Our theory encompasses linear, quadratic, and more general loss functions. We validate the analytic results through numerical simulations. Our findings establish a theoretical foundation for QITE dynamics and provide analytic insights for the first-principle design of variational quantum algorithms.

Analogical reasoning, the transfer of relational structures across contexts (e.g., planet is to sun as electron is to nucleus), is fundamental to scientific discovery. Yet human insight is often constrained by domain expertise and surface-level biases, limiting access to deeper, structure-driven analogies both within and across disciplines. Large language models (LLMs), trained on vast cross-domain data, present a promising yet underexplored tool for analogical reasoning in science. Here, we demonstrate that LLMs can generate novel battery materials by (1) retrieving cross-domain analogs and analogy-guided exemplars to steer exploration beyond conventional dopant substitutions, and (2) constructing in-domain analogical templates from few labeled examples to guide targeted exploitation. These explicit analogical reasoning strategies yield candidates outside established compositional spaces and outperform standard prompting baselines. Our findings position LLMs as interpretable, expert-like hypothesis generators that leverage analogy-driven generalization for scientific innovation.

Prompt engineering, particularly Chain-of-Thought (CoT) prompting, significantly enhances LLM reasoning capabilities. We introduce "Sculpting," a constrained, rule-based prompting method designed to improve upon standard CoT by reducing errors from semantic ambiguity and flawed common sense. We evaluate three prompting strategies (Zero Shot, standard CoT, and Sculpting) across three OpenAI model generations (gpt-4o-mini, gpt-4o, gpt-5) using the GSM8K mathematical reasoning benchmark (1,317 problems). Our findings reveal a "Prompting Inversion": Sculpting provides advantages on gpt-4o (97% vs. 93% for standard CoT), but becomes detrimental on gpt-5 (94.00% vs. 96.36% for CoT on full benchmark). We trace this to a "Guardrail-to-Handcuff" transition where constraints preventing common-sense errors in mid-tier models induce hyper-literalism in advanced models. Our detailed error analysis demonstrates that optimal prompting strategies must co-evolve with model capabilities, suggesting simpler prompts for more capable models.

Luttinger compensated antiferromagnets (LcAFMs), combining spin polarization with vanishing net magnetization, offering distinct advantages for next-generation spintronic applications. Using first-principles calculations, we demonstrate that conventional antiferromagnetic CrCl2 double chains can be transformed into one-dimensional LcAFMs under an external electric field, exhibiting pronounced isotropic spin splitting. The magnitude of the splitting, as well as the band gap, can be effectively tuned by both in-plane and out-of-plane fields, thereby providing greater controllability than in two-dimensional counterparts. To further enhance the tunability, we design a nearly lattice-matched CrCl2/MoTe2 heterostructure and uncover that interfacial charge transfer generates a built-in electric field, inducing spin splitting comparable to that driven by external fields. These results establish interfacial engineering as a highly efficient route to realize and manipulate LcAFM states in low-dimensional magnets, expanding the design principles for spintronic functionalities at the nanoscale.

Metamaterials derive their unconventional properties from engineered microstructures, with periodic lattices providing a versatile framework for modeling wave propagation. Dispersion relations, obtained from Bloch-Floquet theory, govern how waves propagate, attenuate, or localize within such systems. Extending interactions beyond nearest neighbors, through nonlocality, substantially enriches the design space of band diagrams, enabling phenomena such as negative or zero group velocities, roton-like extrema, and band-gap localization. However, existing approaches to dispersion tailoring often rely on analytical formulations or Fourier-based identifications, which become impractical for complex coupling mechanisms and offer limited control over physical constraints such as stiffness positivity. This work introduces a general interpolation-based framework for customizing dispersion relations in uniform nonlocal lattices. Rather than reconstructing full dispersion curves, the method enforces prescribed frequency-wavenumber points as interpolation constraints, enabling localized and tunable control of wave behavior. The formulation is applied to both spring- and beam-interaction lattices, and demonstrated on an Euler-Bernoulli beam model with adjustable nonlocal couplings. Through systematic parameter tuning, the framework enables the creation of rotons, the adjustment of group-velocity dispersion, and the design of evanescent waves with controlled exponential decay within band gaps, all while ensuring real, positive-only stiffness parameters and passive mechanical behavior. Altogether, this parametric interpolation strategy provides a physically consistent and computationally efficient route for engineering advanced phononic functionalities in periodic nonlocal systems.

Safety is a fundamental requirement for autonomous systems operating in critical domains. Control barrier functions (CBFs) have been used to design safety filters that minimally alter nominal controls for such systems to maintain their safety. Learning neural CBFs has been proposed as a data-driven alternative for their computationally expensive optimization-based synthesis. However, it is often the case that the failure set of states that should be avoided is non-obvious or hard to specify formally, e.g., tailgating in autonomous driving, while a set of expert demonstrations that achieve the task and avoid the failure set is easier to generate. We use ICL to train a constraint function that classifies the states of the system under consideration to safe, i.e., belong to a controlled forward invariant set that is disjoint from the unspecified failure set, and unsafe ones, i.e., belong to the complement of that set. We then use that function to label a new set of simulated trajectories to train our neural CBF. We empirically evaluate our approach in four different environments, demonstrating that it outperforms existing baselines and achieves comparable performance to a neural CBF trained with the same data but annotated with ground-truth safety labels.

Generative AI tools are increasingly used to create portrayals of people in occupations, raising concerns about how race and gender are represented. We conducted a large-scale audit of over 1.5 million occupational personas across 41 U.S. occupations, generated by four large language models with different AI safety commitments and countries of origin (U.S., China, France). Compared with Bureau of Labor Statistics data, we find two recurring patterns: systematic shifts, where some groups are consistently under- or overrepresented, and stereotype exaggeration, where existing demographic skews are amplified. On average, White (--31pp) and Black (--9pp) workers are underrepresented, while Hispanic (+17pp) and Asian (+12pp) workers are overrepresented. These distortions can be extreme: for example, across all four models, Housekeepers are portrayed as nearly 100\% Hispanic, while Black workers are erased from many occupations. For HCI, these findings show provider choice materially changes who is visible, motivating model-specific audits and accountable design practices.

Large language models have demonstrated remarkable reasoning capabilities across diverse natural language tasks. However, comparable breakthroughs in scientific discovery are more limited, because understanding complex physical phenomena demands multifaceted representations far beyond language alone. A compelling example is the design of functional materials such as MOFs-critical for a range of impactful applications like carbon capture and hydrogen storage. Navigating their vast and intricate design space in language-based representations interpretable by LLMs is challenging due to the numerous possible three-dimensional atomic arrangements and strict reticular rules of coordination geometry and topology. Despite promising early results in LLM-assisted discovery for simpler materials systems, MOF design remains heavily reliant on tacit human expertise rarely codified in textual information alone. To overcome this barrier, we introduce L2M3OF, the first multimodal LLM for MOFs. L2M3OF integrates crystal representation learning with language understanding to process structural, textual, and knowledge modalities jointly. L2M3OF employs a pre-trained crystal encoder with a lightweight projection layer to compress structural information into a token space, enabling efficient alignment with language instructions. To facilitate training and evaluation, we curate a structure-property-knowledge database of crystalline materials and benchmark L2M3OF against state-of-the-art closed-source LLMs such as GPT-5, Gemini-2.5-Pro and DeepSeek-R1. Experiments show that L2M3OF outperforms leading text-based closed-source LLMs in property prediction and knowledge generation tasks, despite using far fewer parameters. These results highlight the importance of multimodal approaches for porous material understanding and establish L2M3OF as a foundation for next-generation AI systems in materials discovery.

Underground storage of hydrogen and ammonia in geological formations is essential for renewable energy integration, but salt precipitation during gas injection may threaten storage performance. While extensively studied for CO2 systems, precipitation mechanisms in hydrogen-brine and ammonia-brine systems remain poorly understood. This study presents a comprehensive microfluidic investigation of salt crystallization during hydrogen injection into saline and ammonia-containing brines using high-pressure microfluidics. We conducted 81 high-pressure experiments systematically varying brine composition (1-5 mol/kg NaCl), chemical additives (surfactants, alcohols, ammonia), and hydrogen flow rates (200-1300 mL/min). Quantitative image analysis reveals that hydrogen-induced precipitation differs fundamentally from CO2 systems. Hydrogen drives physical precipitation via evaporation and capillary trapping, producing discrete, localized deposits. In contrast, CO2-ammonia systems generate extensive reactive precipitation of ammonium bicarbonate with interconnected crystal networks. Interfacial tension (IFT) controls both residual brine distribution and final crystal coverage: high-IFT fluids form large, interconnected brine pools promoting extensive crystallization, while low-IFT fluids create isolated pools reducing crystal coverage by 50\%. Alcohol and surfactant additives suppress precipitation by enhancing brine mobility, whereas ammonia paradoxically increases crystal fractions despite lower IFT. Higher flow rates accelerate crystallization across all compositions, enabling operational mitigation strategies. and demonstrate that gas-specific, rather than CO2-analog, risk assessments are essential for underground hydrogen storage design. The effectiveness of chemical additives offers promising pathways for near-wellbore protection in underground hydrogen storage operations.

Large language models are often described as capable of reflective reasoning, yet recursive self-evaluation without external feedback frequently yields reformulation rather than progress. We test this prediction in a cross-provider study of 144 reasoning sequences across three models (OpenAI GPT-4o-mini, Anthropic Claude 3 Haiku, and Google Gemini 2.0 Flash) and four task families (arithmetic, code, explanation, reflection), each iterated ten times under two conditions: ungrounded self-critique and a minimal grounding intervention (a single verification step at iteration three). Mean informational change (delta I, measured via normalized edit distance) declined by 55% from early (0.193) to late (0.087) iterations in ungrounded runs, with consistent patterns across all three providers. Grounded runs showed a +28% rebound in informational change immediately after the intervention and sustained non-zero variance thereafter. Complementary measures-n-gram novelty, embedding drift, and character-level entropy-converged on the same pattern: reflection without contact tends toward informational closure. We interpret this as evidence for a structural limit on self-correction in generative reasoning: without an exchange of information with an independent verifier or environment, recursive inference approaches an attractor state of epistemic stasis. Minimal grounding functions as dissipative coupling, reintroducing informational flux. The cross-architecture consistency suggests the mirror loop arises from shared autoregressive training objectives rather than provider-specific alignment schemes. The results delineate when reflection is performative rather than epistemic and motivate design principles for grounded, cooperative reasoning. Materials and code are publicly available.

Magnetic topology and its associated emergent phenomena are central to realizing intriguing quantum states and spintronics functionalities. Designing spin textures to achieve strong and distinct electrical responses remains a significant challenge. Layered transition metal dichalcogenides offer a versatile platform for tailoring structural and magnetic properties, enabling access to a wide spectrum of topological magnetic states. Here, we report a domain-wall-driven, large, and tunable topological Hall effect (THE) in a non-centrosymmetric intercalated transition metal dichalcogenides series Fe$_x$TaS$_2$. By systematically varying the Fe intercalation level, we exert precise control over the magnetic ground states, allowing manipulation of the topological Hall effect. Real-space magnetic force microscopy (MFM) provides direct evidence of periodic magnetic stripe domain formation, confirming the microscopic origin of the observed topological transport phenomena. Our findings establish a promising way for tuning the topology of domains to generate substantial electromagnetic responses in layered magnetic materials.

Powder X-ray diffraction (XRD) is a foundational technique for characterizing crystalline materials. However, the reliable interpretation of XRD patterns, particularly in multiphase systems, remains a manual and expertise-demanding task. As a characterization method that only provides structural information, multiple reference phases can often be fit to a single pattern, leading to potential misinterpretation when alternative solutions are overlooked. To ease humans' efforts and address the challenge, we introduce Dara (Data-driven Automated Rietveld Analysis), a framework designed to automate the robust identification and refinement of multiple phases from powder XRD data. Dara performs an exhaustive tree search over all plausible phase combinations within a given chemical space and validates each hypothesis using a robust Rietveld refinement routine (BGMN). Key features include structural database filtering, automatic clustering of isostructural phases during tree expansion, peak-matching-based scoring to identify promising phases for refinement. When ambiguity exists, Dara generates multiple hypothesis which can then be decided between by human experts or with further characteriztion tools. By enhancing the reliability and accuracy of phase identification, Dara enables scalable analysis of realistic complex XRD patterns and provides a foundation for integration into multimodal characterization workflows, moving toward fully self-driving materials discovery.

Powder X-ray diffraction (XRD) is a foundational technique for characterizing crystalline materials. However, the reliable interpretation of XRD patterns, particularly in multiphase systems, remains a manual and expertise-demanding task. As a characterization method that only provides structural information, multiple reference phases can often be fit to a single pattern, leading to potential misinterpretation when alternative solutions are overlooked. To ease humans' efforts and address the challenge, we introduce Dara (Data-driven Automated Rietveld Analysis), a framework designed to automate the robust identification and refinement of multiple phases from powder XRD data. Dara performs an exhaustive tree search over all plausible phase combinations within a given chemical space and validates each hypothesis using a robust Rietveld refinement routine (BGMN). Key features include structural database filtering, automatic clustering of isostructural phases during tree expansion, peak-matching-based scoring to identify promising phases for refinement. When ambiguity exists, Dara generates multiple hypothesis which can then be decided between by human experts or with further characteriztion tools. By enhancing the reliability and accuracy of phase identification, Dara enables scalable analysis of realistic complex XRD patterns and provides a foundation for integration into multimodal characterization workflows, moving toward fully self-driving materials discovery.

This pilot study traces students' reflections on the use of AI in a 13-week foundational design course enrolling over 500 first-year engineering and architecture students at the Singapore University of Technology and Design. The course was an AI-enhanced design course, with several interventions to equip students with AI based design skills. Students were required to reflect on whether the technology was used as a tool (instrumental assistant), a teammate (collaborative partner), or neither (deliberate non-use). By foregrounding this three-way lens, students learned to use AI for innovation rather than just automation and to reflect on agency, ethics, and context rather than on prompt crafting alone. Evidence stems from coursework artefacts: thirteen structured reflection spreadsheets and eight illustrated briefs submitted, combined with notes of teachers and researchers. Qualitative coding of these materials reveals shared practices brought about through the inclusion of Gen-AI, including accelerated prototyping, rapid skill acquisition, iterative prompt refinement, purposeful "switch-offs" during user research, and emergent routines for recognizing hallucinations. Unexpectedly, students not only harnessed Gen-AI for speed but (enabled by the tool-teammate-neither triage) also learned to reject its outputs, invent their own hallucination fire-drills, and divert the reclaimed hours into deeper user research, thereby transforming efficiency into innovation. The implications of the approach we explore shows that: we can transform AI uptake into an assessable design habit; that rewarding selective non-use cultivates hallucination-aware workflows; and, practically, that a coordinated bundle of tool access, reflection, role tagging, and public recognition through competition awards allows AI based innovation in education to scale without compromising accountability.

Predicting which hypothetical inorganic crystals can be experimentally realized remains a central challenge in accelerating materials discovery. SyntheFormer is a positive-unlabeled framework that learns synthesizability directly from crystal structure, combining a Fourier-transformed crystal periodicity (FTCP) representation with hierarchical feature extraction, Random-Forest feature selection, and a compact deep MLP classifier. The model is trained on historical data from 2011 through 2018 and evaluated prospectively on future years from 2019 to 2025, where the positive class constitutes only 1.02 per cent of samples. Under this temporally separated evaluation, SyntheFormer achieves a test area under the ROC curve of 0.735 and, with dual-threshold calibration, attains high-recall screening with 97.6 per cent recall at 94.2 per cent coverage, which minimizes missed opportunities while preserving discriminative power. Crucially, the model recovers experimentally confirmed metastable compounds that lie far from the convex hull and simultaneously assigns low scores to many thermodynamically stable yet unsynthesized candidates, demonstrating that stability alone is insufficient to predict experimental attainability. By aligning structure-aware representation with uncertainty-aware decision rules, SyntheFormer provides a practical route to prioritize synthesis targets and focus laboratory effort on the most promising new inorganic materials.

Traditional machine learning has advanced polymer discovery, yet direct generation of chemically valid and synthesizable polymers without exhaustive enumeration remains a challenge. Here we present polyT5, an encoder-decoder chemical language model based on the T5 architecture, trained to understand and generate polymer structures. polyT5 enables both property prediction and the targeted generation of polymers conditioned on desired property values. We demonstrate its utility for dielectric polymer design, seeking candidates with dielectric constant >3, bandgap >4 eV, and glass transition temperature >400 K, alongside melt-processability and solubility requirements. From over 20,000 generated promising candidates, one was experimentally synthesized and validated, showing strong agreement with predictions. To further enhance usability, we integrated polyT5 within an agentic AI framework that couples it with a general-purpose LLM, allowing natural language interaction for property prediction and generative design. Together, these advances establish a versatile and accessible framework for accelerated polymer discovery.

We employ interpretable explicit machine learning to analyze the material dependence of the magnetic transition temperature $T_c$ in ferromagnetic and ferrimagnetic Heusler compounds. For around 200 compounds, we consider both experimental $T_c$ and calculated $T_c$ using \textit{ab initio} determination of magnetic interactions together with a Monte-Carlo solution. We use the hierarchical dependence extraction (HDE) procedure [Mor\'ee and Arita, Phys. Rev. B 110, 014502 (2024)] to extract the dependencies of $T_c$ on chemical proportions and magnetic moments from the main order to the higher order, and construct an explicit expression of $T_c$ from these dependencies. The main results are: (a) $T_c$ is mainly controlled by the proportions of Fe, Co, and Mn, and increases with these proportions, consistent with previous machine learning analyses of ferromagnetic materials. (b) The HDE describes $T_c$ with an accuracy that is comparable to that of other machine learning procedures. (c) The HDE expression of $T_c$ can be interpreted as a generalized order parameter that increases with increasing magnetization amplitude, in qualitative agreement with various theories of phase transitions. These results strengthen our understanding of the material dependence of $T_c$ in collinear Heusler magnets and motivate the further use of HDE in material design.

Machine learning has accelerated materials discovery, yet most high-performing models remain black boxes, offering predictions without physical understanding. Here I present GoodRegressor, a general-purpose, C++-based symbolic regression framework that bridges data-driven modeling and physical interpretability. GoodRegressor systematically explores nonlinear transformations and feature interactions across five integrated modules, parser, designer, curator, regressor, and designer as a post-process, to construct compact, physics-consistent analytical models. For example, applied to the experimental activation energy dataset of oxygen-ion conductors, GoodRegressor explored an ensemble-sampled model space comprising approximately $1.44 \times 10^{457}$ possible combinations and achieved superior predictive performance ($\langle R^2 \rangle =0.804$) compared with conventional machine learning methods, outperforming RandomForest, XGBoost, LightGBM, Ridge, MLP, and PySR ($\langle R^2 \rangle \leq 0.652$). Unlike black-box models, GoodRegressor reveals transparent structure-property relationships linking ionic transport to coordination environment and lattice flexibility. This interpretable modeling framework mitigates the opacity of conventional ML, enabling hypothesis generation, physical insight, and general applicability to complex scientific systems beyond materials informatics.

Symbolic regression offers a promising route toward interpretable machine learning, yet existing methods suffer from poor predictability and computational intractability when exploring large expression spaces. I introduce GoodRegressor, a general-purpose C++-based framework that resolves these limitations while preserving full physical interpretability. By combining hierarchical descriptor construction, interaction discovery, nonlinear transformations, statistically rigorous model selection, and stacking ensemble, GoodRegressor efficiently explores symbolic model spaces such as $1.44 \times 10^{457}$, $5.99 \times 10^{124}$, and $4.20 \times 10^{430}$ possible expressions for oxygen-ion conductors, NASICONs, and superconducting oxides, respectively. Across these systems, it produces compact equations that surpass state-of-the-art black-box models and symbolic regressors, improving $R^2$ by $4 \sim 40$ %. The resulting expressions reveal physical insights, for example, into oxygen-ion transport through coordination environment and lattice flexibility. Independent ensemble runs yield nearly identical regressed values and the identical top-ranked candidate, demonstrating high reproducibility. With scalability up to $10^{4392}$ choices without interaction terms, GoodRegressor provides a foundation for general-purpose interpretable machine intelligence.

Artificial intelligence (AI) is influencing heterogeneous catalysis research by accelerating simulations and materials discovery. A key frontier is integrating AI with multiscale models and multimodal experiments to address the "many-to-one" challenge of linking intrinsic kinetics to observables. Advances in machine-learned force fields, microkinetics, and reactor modeling enable rapid exploration of chemical spaces, while operando and transient data provide unprecedented insight. Yet, inconsistent data quality and model complexity limit mechanistic discovery. Generative and agentic AI can automate model generation, quantify uncertainty, and couple theory with experiment, realizing "self-driving models" that produce interpretable, reproducible, and transferable understanding of catalytic systems.

We present a comprehensive theoretical investigation of magnetoelectric (ME) coupling mechanisms in 19 altermagnetic and 4 ferrimagnetic Type-I multiferroics using electronic band structure calculations with spin-orbit coupling, a first-principles ME response framework, and spin-space-group theory analysis. We formulate a universal scheme for realizing nonvolatile ME coupling in Type-I multiferroics, where two distinct pathways emerge, each dictated by spin-space symmetry. The first pathway is associated with switching of the spin splitting or the now familiar spin-momentum locking in reciprocal space, characteristic of some altermagnetic mul-tiferroics that exhibit coexisting antiferromagnetism and ferroelectricity. The second pathway involves real-space magnetization switching via electric polarization reversal, characterized by switchable components of the linear ME tensor, despite the traditionally weak coupling in Type-I systems due to the independent origins of magnetism and ferroelectricity. We demonstrate that these two intrinsic ME coupling mechanisms are mutually exclusive and propose thermodynami-cally stable compounds for experimentation. Our findings establish general design principles for controlling robust nonvolatile ME effects in multiferroic materials.

This work reports optical and electronic numerical modelling of a novel emerging structure which is the GaAs nanocrystal on Si tandem solar cell by epitaxial lateral overgrowth, a technique which allows defect free material growth. The techniqueconsists of creating nucleation sites in a silicon surface SiO2 layer and initiating growth of nanoscalescale seeds, whereby strain energy remains below the Matthews-Blakeslee strain relaxation limit. This leads to AlxGaAs growth in micro-crystals without generation of material defects. The focus of this presentation is optical and electrical modelling of nanocrystals for applications in the very active field of silicon based multijunction solar cells, and design of a AlxGaAs/Si two terminal tandem, for compositions ranging from x=0 to x=30% in absorber layers. We present a model of the complete structure in two dimensions, consisting of a Al xGaAs high bandgap subcell connected with a tunnel junction to the low bandgap Si junction. The elaboration of models is described, with an emphasis on the AlxGaAs crystal featuring a non-planar pn-junction, and a focus on the optical properties of this lattice of micrometric AlGaAs crystals and in particular their light trapping properties from the resulting surface texture. The question of AlxGaAs surface coverage is addressed, given that neighbouring AlxGaAs crystals have different crystal orientations on a (111) Si surface, such that any coalescence of neighbour AlxGaAs crystals leads to crippling defects at their interface. The result is that some high energy incident light above the AlxGaAs bandgap is nevertheless transmitted directly to the Si cell, such that the resulting photogenerated carriers thermalise to the Silicon bandgap, and result in a loss of efficiency. The interface between AlxGaAs and Si subcells is addressed, with an emphasis on current transport efficiency through the nanoseeds and tunnelling currents through appropriately designed SiO2 buffer layers. This work therefore presents a theoretical framework for evaluating the potential of AlxGaAs nanocrystal growth on Si for light trapping, for GaAs silicon two terminal tandem cell performance including tunnel junctions, and provides models and design rules for efficient AlxGaAs microcrystal arrays as high bandgap subcells for tandem solar cells on silicon.

In the same way that generative models today conduct most of their training in a self-supervised fashion, how can agentic models conduct their training in a self-supervised fashion, interactively exploring, learning, and preparing to quickly adapt to new tasks? A prerequisite for embodied agents deployed in real world interactions ought to be training with interaction, yet today's most successful AI models (e.g., VLMs, LLMs) are trained without an explicit notion of action. The problem of pure exploration (which assumes no data as input) is well studied in the reinforcement learning literature and provides agents with a wide array of experiences, yet it fails to prepare them for rapid adaptation to new tasks. Today's language and vision models are trained on data provided by humans, which provides a strong inductive bias for the sorts of tasks that the model will have to solve (e.g., modeling chords in a song, phrases in a sonnet, sentences in a medical record). However, when they are prompted to solve a new task, there is a faulty tacit assumption that humans spend most of their time in the most rewarding states. The key contribution of our paper is a method for pre-training interactive agents in a self-supervised fashion, so that they can instantly mimic human demonstrations. Our method treats goals (i.e., observations) as the atomic construct. During training, our method automatically proposes goals and practices reaching them, building off prior work in reinforcement learning exploration. During evaluation, our method solves an (amortized) inverse reinforcement learning problem to explain demonstrations as optimal goal-reaching behavior. Experiments on standard benchmarks (not designed for goal-reaching) show that our approach outperforms prior methods for zero-shot imitation.

Rare earth doping is an effective way to convert chemically stable oxides into multifunctional materials with coupled electronic, optical, and magnetic properties. We present first principles calculations of pristine and Tm3+ doped Ca2SnO4 to understand how localized 4f states change the structural, electronic, magnetic, and optical behavior of the host. Pristine Ca2SnO4 is a mechanically stable, wide band gap insulator with mostly ionic covalent bonding and diamagnetic character. Replacing Ca2+ with Tm3+ introduces several key changes: (i) localized Tm 4f states create intermediate levels inside the wide gap, reducing the optical band gap; (ii) exchange and spin orbit interactions generate strong local magnetic moments and spin asymmetry near the conduction band; (iii) electron localization function analysis shows enhanced covalency and electron pockets that stabilize luminescent centers; and (iv) the optical response shows visible range absorption, refractive index features, and low energy plasmon peaks while maintaining high energy dielectric stability. These effects make Tm doped Ca2SnO4 a mechanically robust, optically tunable, and magnetically active oxide phosphor suitable for red emission, intermediate band photovoltaics, and spin photon coupling. More broadly, our results show how targeted rare earth substitution can enable multifunctionality in wide gap stannates and guide the design of next generation spintronic photonic oxides.

In this work, we propose a sparse transformer architecture that incorporates prior information about the underlying data distribution directly into the transformer structure of the neural network. The design of the model is motivated by a special optimal transport problem, namely the regularized Wasserstein proximal operator, which admits a closed-form solution and turns out to be a special representation of transformer architectures. Compared with classical flow-based models, the proposed approach improves the convexity properties of the optimization problem and promotes sparsity in the generated samples. Through both theoretical analysis and numerical experiments, including applications in generative modeling and Bayesian inverse problems, we demonstrate that the sparse transformer achieves higher accuracy and faster convergence to the target distribution than classical neural ODE-based methods.

Pi-conjugated covalent organic frameworks (COFs) provide a versatile platform for the realization of designer quantum nanomaterials. Strong electron-electron correlation within these artificial lattices can give rise to exotic phases of matter. Their experimental realization however requires precise control over orbital symmetry, charge localization, and band dispersion all arising from the effective hybridization between molecular linkers and nodes. Here, we present a modular strategy for constructing diatomic Kagome lattices from aza-[3]triangulene (A[3]T) nodes, in which a D3h symmetric ground state is stabilized through resonance contributions from a cumulenenic linker. First-principles density-functional theory and scanning tunnelling spectroscopy reveal that the hybridization of a sixfold degenerate set of edge-localized Wannier functions in the unit cell gives rise to orbital-phase frustration-induced non-trivial flat bands. These results establish a general design principle for engineering orbital interactions in organic lattices and open a pathway toward programmable COF-based quantum materials with correlated electronic ground states.

Abnormal grain growth (AGG) influences the properties of polycrystalline materials; however, the underlying mechanisms, particularly the role of solute segregation at the grain boundary (GB), are difficult to quantify precisely. This study demonstrates a multiscale framework that integrates atomic-scale segregation energetics (using density functional theory) with mesoscale grain growth dynamics (using phase-field model) to investigate AGG, using $α$-Fe as an example system. Multisite segregation energies are calculated for symmetric tilt grain boundaries (STGBs) along the $\langle 110 \rangle$ axis for nine different solutes (Co, Cr, Mn, Mo, Nb, Ni, Ti, W, and V), encompassing three different types of coincident site lattice (CSL) boundaries: $\sum 3 (11\bar{2})$, $\sum 9 (\bar{2}21)$, and $\sum 3 (\bar{1}11)$. The model takes into account the effect of solute drag on GB mobility, estimated using a bulk solute concentration of 0.1 at\%. The results demonstrate that AGG originates due to GB anisotropy, the extent of which largely depends on the type of solute atom present. Such a complex dependence necessitates using a multiscale model to understand AGG comprehensively. In general, low-energy $Σ3$ boundaries are found to have higher mobility and show preferential growth for most of the solutes, other than Co. The study reveals how the distribution of GB types significantly influences AGG. When 10-30\% of the GBs are high-mobility type, crown-like morphologies are observed, leading to AGG. These findings underscore the critical role of GB chemistry and crystallography in governing AGG, and the model can be generalized to provide a predictive framework for controlling grain growth through strategic solute design in advanced alloys.

Accurate prediction of ionic conductivity in electrolyte systems is crucial for advancing numerous scientific and technological applications. While significant progress has been made, current research faces two fundamental challenges: (1) the lack of high-quality standardized benchmarks, and (2) inadequate modeling of geometric structure and intermolecular interactions in mixture systems. To address these limitations, we first reorganize and enhance the CALiSol and DiffMix electrolyte datasets by incorporating geometric graph representations of molecules. We then propose GeoMix, a novel geometry-aware framework that preserves Set-SE(3) equivariance-an essential but challenging property for mixture systems. At the heart of GeoMix lies the Geometric Interaction Network (GIN), an equivariant module specifically designed for intermolecular geometric message passing. Comprehensive experiments demonstrate that GeoMix consistently outperforms diverse baselines (including MLPs, GNNs, and geometric GNNs) across both datasets, validating the importance of cross-molecular geometric interactions and equivariant message passing for accurate property prediction. This work not only establishes new benchmarks for electrolyte research but also provides a general geometric learning framework that advances modeling of mixture systems in energy materials, pharmaceutical development, and beyond.

Understanding and optimizing polysulfide adsorption and conversion processes are critical to mitigating shuttle effects and sluggish redox kinetics in lithium-sulfur batteries (LSBs). Here, we introduce a machine-learning-accelerated framework, Precise and Accurate Configuration Evaluation (PACE), that integrates Machine Learning Interatomic Potentials (MLIPs) with Density Functional Theory (DFT) to systematically explore adsorption configurations and energetics of a series of N6-coordinated dual-atom catalysts (DACs). Our results demonstrate that, compared with single-atom catalysts, DACs exhibit improved LiPS adsorption and redox conversion through cooperative metal-sulfur interactions and electronic coupling between adjacent metal centers. Among all DACs, Fe-Ni and Fe-Pt show optimal catalytic performance, due to their optimal adsorption energies (-1.0 to -2.3 eV), low free-energy barriers (<=0.4 eV) for the Li2S2 to Li2S conversion, and facile Li2S decomposition barriers (<=1.0 eV). To accelerate catalyst screening, we further developed a machine learning (ML) regression model trained on DFT-calculated data to predict the Gibbs free energy (ΔG) of Li2Sn adsorption using physically interpretable descriptors. The Gradient Boosting Regression (GBR) model yields an R^2 of 0.85 and an MAE of 0.26 eV, enabling the rapid prediction of ΔG for unexplored DACs. Electronic-structure analyses reveal that the superior performance originates from the optimal d-band alignment and S-S bond polarization induced by the cooperative effect of dual metal centres. This dual ML-DFT framework demonstrates a generalizable, data-driven design strategy for the rational discovery of efficient catalysts for next-generation LSBs.

The photocatalytic water-splitting process is thermodynamically challenging and requires catalysts with suitable band structures, as well as the presence of supporting cocatalysts. By considering the unique charge carrier mobility in perovskites, this study introduces three new ABO3-type high-entropy perovskites (Ba1/2Sr1/2)(Ti1/3Zr1/3Hf1/3)O3, (Ba1/2Sr1/2)(Ga1/3In1/3Sn1/3)O3 and (Ba1/2Sr1/2)(Ti1/3Zr1/3Sn1/3)O3 for cocatalyst-free photocatalysis. The three catalysts, having a single-phase cubic structure, are designed by considering configurational entropy, tolerance factor, octahedral factor, ionic radius deviation and valence deviation of >1.5R (R: gas constant), 0.9-1.0, 0.4-0.8, >0.3 and >0.3, respectively. The perovskites exhibit similar valence band tops, while their bandgaps vary slightly depending on the composition at the B-site (slightly lower bandgap by including d10 cations). Additionally, all three materials demonstrate effective hydrogen generation without the need for added cocatalysts. This investigation confirms that high-entropy oxide perovskites can offer significant potential for cocatalyst-free photocatalytic reactions.

Additive manufacturing (AM) processes, particularly laser powder bed fusion, offer exceptional design flexibility but impose extreme thermal conditions that limit the processability of high strength aluminum alloys. This work develops an Integrated Computational Materials Engineering (ICME) framework that combines physics informed machine learning (ML) with large scale molecular dynamics (MD) simulations to enable inverse design of Al alloys with improved strength ductility synergy tailored for AM. A heterogeneous multi process database of about 1500 alloys, spanning laser based and solid state AM, cast, and wrought routes, was assembled with yield strength, ultimate tensile strength, and elongation as target properties. Composition derived and rule of mixtures descriptors created from elemental properties define a 46 dimensional, physically interpretable feature space. A Random Forest regressor provides strong performance for strength (R2 about 0.85) and competitive accuracy for ductility, while SHAP analysis highlights stacking fault energy, solute misfit, and thermophysical descriptors as dominant factors. Robustness tests using fractional unknown row validation and evaluation on recently reported 2024 2025 AM alloys confirm that the model generalizes across composition process space. A strength ductility quality index is then used to map design spaces in benchmark alloy families, including binary Al Cu, quaternary Al Zn Mg Cu, and the Al92(Ti,Fe,Co,Ni)2 family, consistently identifying narrow composition windows with high synergy. Large scale MD simulations of Al(Fe,Co,Ni,Ti) alloys with varying Al content reproduce an inverted U shaped dependence of strength and ductility with a maximum near Al88 to Al92, in close agreement with ML predictions. The results demonstrate a data driven, physics aware workflow for accelerating discovery of next generation aluminum alloys for AM.

Camera movement conveys spatial and narrative information essential for understanding video content. While recent camera movement classification (CMC) methods perform well on modern datasets, their generalization to historical footage remains unexplored. This paper presents the first systematic evaluation of deep video CMC models on archival film material. We summarize representative methods and datasets, highlighting differences in model design and label definitions. Five standard video classification models are assessed on the HISTORIAN dataset, which includes expert-annotated World War II footage. The best-performing model, Video Swin Transformer, achieves 80.25% accuracy, showing strong convergence despite limited training data. Our findings highlight the challenges and potential of adapting existing models to low-quality video and motivate future work combining diverse input modalities and temporal architectures.

We assess the performance of EUCLID, Efficient Unsupervised Constitutive Law Identification and Discovery, a recently proposed framework for automated discovery of constitutive laws, on experimental data. Mechanical tests are performed on natural rubber specimens spanning simple to complex geometries, from which we collect both global, force elongation, and local, full-field displacement, measurements. Using these data, we obtain constitutive laws via two routes, the conventional identification of unknown parameters in a priori selected material models, and EUCLID, which automates model selection and parameter identification within a unified model-discovery pipeline. We compare the two methodologies using global versus local data, analyze predictive accuracy, and examine generalization to unseen geometries. Moreover, we quantify the experimental noise, investigate the coverage of the material state space achieved by each approach and discuss the relative performance of different datasets and different a priori chosen models versus EUCLID.

Polymers, macromolecules formed from covalently bonded monomers, underpin countless technologies and are indispensable to modern life. While deep learning is advancing polymer science, existing methods typically represent the whole polymer solely through monomer-level descriptors, overlooking the global structural information inherent in polymer conformations, which ultimately limits their practical performance. Moreover, this field still lacks a universal foundation model that can effectively support diverse downstream tasks, thereby severely constraining progress. To address these challenges, we introduce PolyConFM, the first polymer foundation model that unifies polymer modeling and design through conformation-centric generative pretraining. Recognizing that each polymer conformation can be decomposed into a sequence of local conformations (i.e., those of its repeating units), we pretrain PolyConFM under the conditional generation paradigm, reconstructing these local conformations via masked autoregressive (MAR) modeling and further generating their orientation transformations to recover the corresponding polymer conformation. Besides, we construct the first high-quality polymer conformation dataset via molecular dynamics simulations to mitigate data sparsity, thereby enabling conformation-centric pretraining. Experiments demonstrate that PolyConFM consistently outperforms representative task-specific methods on diverse downstream tasks, equipping polymer science with a universal and powerful tool.

Two-dimensional (2D) materials with a twist between layers exhibit a moir\'e interference pattern with larger periodicity than any of the constituent layer unit cells. In these systems, a wealth of exotic phases appear that result from moir\'e-dependent many-body electron correlation effects or non-trivial band topology. One problem with using twist to generate moir\'e interference has been the difficulty in creating high-quality, uniform, and repeatable samples due to fabrication through mechanical stacking with viscoelastic stamps. Here we show, a new method to generate moir\'e interference through the controlled application of layer-by-layer strain (heterostrain) on non-twisted 2D materials, where moir\'e interference results from strain-induced lattice mismatch without twisting or stacking. Heterostrain generation is achieved by depositing stressed thin films onto 2D materials to apply large strains to the top layers while leaving layers further down less strained. We achieve deterministic control of moir\'e periodicity and symmetry in non-twisted 2D multilayers and bilayers, with 97% yield, through varying stressor film force (film thickness X film stress) and geometry. Moir\'e reconstruction effects are memorized after the removal of the stressor layers. Control over the strain degree-of-freedom opens the door to a completely unexplored set of unrealized tunable moir\'e geometric symmetries, which may now be achieved in a high-yield and user-skill independent process taking only hours. This technique solves a long-standing throughput bottleneck in new moir\'e quantum materials discovery and opens the door to industrially-compatible manufacturing for 2D moir\'e-based electronic or optical devices.

Benchmarking outcomes increasingly govern trust, selection, and deployment of LLMs, yet these evaluations remain vulnerable to semantically equivalent adversarial perturbations. Prior work on adversarial robustness in NLP has emphasized text attacks that affect many models equally, leaving open the question of whether it is possible to selectively degrade or enhance performance while minimally affecting other models. We formalize this problem and study selective adversarial attacks on MMLU - a widely used benchmark designed to measure a language model's broad general knowledge and reasoning ability across different subjects. Using canonical attacks integrated into TextAttack framework, we introduce a protocol for selectivity assessment, develop a custom constraint to increase selectivity of attacks and propose a surrogate-LLM pipeline that generates selective perturbations. Empirically, we find that selective adversarial attacks exist and can materially alter relative rankings, challenging the fairness, reproducibility, and transparency of leaderboard-driven evaluation. Our results motivate perturbation-aware reporting and robustness diagnostics for LLM evaluation and demonstrate that even subtle edits can shift comparative judgments.

Two dimensional Ruddlesden Popper (2D) RP hybrid perovskites exhibit substantially higher chemical and structural stability than their three dimensional (3D) counterparts, positioning them as promising candidates for next generation optoelectronics. While quasiparticle dynamics in 3D perovskites are well studied, their 2D analogues remain comparatively underexplored. Here we systematically investigate the branching, dynamics, and interactions of free excitons (FEs) and exciton polarons EPs in monolayer 2D RP perovskites using visible range femtosecond transient absorption TA spectroscopy. We prepared monolayer 2D RP perovskite thin films with varied organic spacers and distinct fabrication routes for comparative analysis. We find that the EP binding energy is 50 65 meV in (BA)2PbI4 and 37 39 meV in (PEA)2PbI4, consistent with spacer layer dependent coupling as corroborated by FTIR. We reveal a dynamic equilibrium between FEs and EPs that persists for tens of picoseconds. Notably, the TA signatures differ by fabrication route films from the newly developed process show weaker Auger annihilation and a reduced hot phonon bottleneck than those from the conventional route trends consistent with fewer traps and impurities in the former. Coupled rate equation modeling reproduces the transients and quantifies the processes of hot carrier relaxation, exciton exciton annihilation, exciton phonon coupling, and FE EP interconversion. These results demonstrate that the chemical synthetic process (fabrication route) and spacer choice significantly influence EP stability and population balance, offering practical levers for engineering ultrafast photophysics in 2D perovskites and guiding the design of advanced optoelectronic devices.

Valley Hall effect is fundamental to valleytronics and provides a promising avenue for advancing information technology. While conventional valley Hall effect requires the inversion symmetry breaking, the recently proposed nonlinear valley Hall (NVH) effect removes the symmetry constraint, and broaden material choices. However, existing studies are limited to nonmagnetic materials without spin involvement and rely on external strain to break rotational symmetry. Here, to address these limitations, we design a magnetically controllable NVH effect in centrosymmetric ferromagnets, by the tight-binding model and first-principles calculations. The model calculations demonstrate nonvanishing NVH conductivities can emerge in pristine hexagonal lattice without external strain, with the magnitude, sign, and spin polarization of the conductivities being all dependent on the magnetization orientation. The effect thus generates various spin-polarized valley Hall currents, characterized by distinct combinations of current direction and spin polarization. First-principle results on a ferromagnetic VSi$_2$N$_4$ bilayer confirm considerable NVH conductivities and their dependence on the magnetization. The magnetically controllable NVH effect unlocks the potential of centrosymmetric magnets for valleytronics, and offer opportunities for novel spintronic and valleytronic devices.

Two-dimensional (2D) nitride materials such as hexagonal boron nitride (h-BN), graphitic carbon nitride (g-C$_3$N$_4$), and beryllonitrene (BeN$_4$) have emerged as promising candidates for next generation electronic, optoelectronic, and energy applications due to their unique structural and electronic properties. This study presents a systematic investigation of the effects of vacancy defect, specifically the role of nitrogen and constituent atom vacancies on the electronic properties of these materials. Our findings reveal that the introduction of nitrogen vacancies significantly alters the electronic characteristics of these materials. In h-BN, the presence of a nitrogen monovacancy significantly lowers the work function from 5.97 eV to 3.45 eV, one of the lowest values reported for any 2D material. Additionally, this defect reduces the band gap from 4.6 eV to 0.64 eV, driving the material toward half-metallic behavior. This is accompanied by the emergence of flat bands near the Fermi level, indicative of strong electron-electron interactions. In g-C$_3$N$_4$, nitrogen vacancies lead to a decrease in work function and band gap, with double nitrogen vacancies rendering the material nearly metallic. In BeN$_4$, nitrogen vacancies result in minimal charge redistribution and a slight increase in work function, highlighting the material's unique electronic behavior. These results underscore the potential of vacancy engineering in tuning the electronic properties of 2D nitride materials, offering avenues for the design of materials with tailored work functions and band gaps for applications in optoelectronics, spintronics, and catalysis.

Recent progress in spintronics within the paradigm of altermagnets (AMs) opens new avenues for next-generation electronic device design. Here, we establish a spin-corner locking mechanism that generates second-order topological states in two-dimensional (2D) altermagnetic systems, through effective model analysis. Remarkably, the breaking of Mxy symmetry under uniaxial strain creates spin-resolved corner modes, driving the system into a corner-polarized second-order topological insulator (CPSOTI). Beyond critical strain, a topological phase transition to quantum anomalous Hall insulator occurs with quantized conductance. Through first-principles calculations, we identify two experimentally viable candidates for 2D intrinsic AM CrO and Cr$_2$Se$_2$O -- which host robust CPSOTI. Moreover, we construct the topological phase diagram of CrO and predict the existence of an altermagnetic Weyl semimetal phase. Our findings open technological avenues in altermagnetism and higher-order topology, while providing opportunities for coupling topological spintronics with cornertronics.

The quest for robust, intrinsically magnetic topological materials exhibiting the quantum anomalous Hall (QAH) effect is a central challenge in condensed matter physics and the application of revolutionary electronics. However, progress has been hampered by the limited number of candidate materials, which often suffer from poor stability and complex synthesis. Here, we introduce a new paradigm by exploring the emergent magnetism and nontrivial band topology in the largely overlooked family of two-dimensional (2D) pentagonal MN$_8$ monolayers. Employing first-principles calculations, we reveal that these systems host out-of-plane ferromagnetic ground states, a key feature that unlocks nontrivial topological properties driven by the localized $d$-orbitals of the embedded transition metals. Remarkably, we identify TiN$_8$ as a QAH insulator characterized by a Chern number of $C=-1$. Even more strikingly, MoN$_8$ is predicted to be a rare high-Chern-number QAH insulator, boasting a Chern number of $C=2$. Our findings establish the penta-MN$_8$ family as a fertile and versatile platform for realizing exotic topological quantum states. This work not only significantly expands the material landscape for magnetic topological insulators but also provides a solid theoretical foundation for designing next-generation spintronic and quantum computing devices.

Organic - inorganic heterostructures (HS) combine the strong light absorption and exciton generation capabilities of organic molecules with the unique excitonic properties of layered transition metal dichalcogenides (TMDs), where the interfacial band alignment dictates the optical response. In this work, we investigate the influence of H2Pc molecules on CVD-grown MoS2 and WS2 monolayers using correlative microscopy techniques - Kelvin probe force microscopy (KPFM), photoluminescence (PL), and Raman spectroscopy. Comprehensive analysis of both electronic and optical properties provides detailed insights into the energy band alignment in these two HS. Despite their similar band alignments, the heterostructures exhibit strikingly different optical signatures. In the case of H2Pc/MoS2 HS, the effect of defect healing is more pronounced, while for the H2Pc/WS2 HS, strong indications of Förster energy transfer are observed. These findings highlight the critical role of transition dipole moment in addition to spectral overlap between donor emission and acceptor absorption in the design of optoelectronic devices.

Efficient exploration of the vast chemical space is a fundamental challenge in materials discovery, particularly for designing functional inorganic crystalline materials with targeted properties. Diffusion-based generative models have emerged as a powerful route, but most existing approaches require domain-specific constraints and separate diffusion processes for atom types, atomic positions, and lattice parameters, adding complexity and limiting efficiency. Here, we present DiffCrysGen, a fully data-driven, score-based diffusion model that generates complete crystal structures in a single, end-to-end diffusion process. This unified framework simplifies the model architecture and accelerates sampling by two to three orders of magnitude compared to existing methods without compromising chemical and structural diversity of the generated materials. In order to demonstrate the efficacy of DiffCrysGen in generating valid and useful materials, using density functional theory (DFT), we validate a number of newly generated rare earth-free magnetic materials that are energetically and dynamically stable, and are potentially synthesizable. These include ferromagnets with high saturation magnetization and large magnetocrystalline anisotropy, as also metallic antiferromagnets. These results establish DiffCrysGen as a general platform for accelerated functional materials discovery.

Efficient exploration of the vast chemical space is a fundamental challenge in materials design and discovery, particularly for designing functional inorganic crystalline materials with targeted properties. Diffusion-based generative models have emerged as a powerful route, but most existing approaches require domain-specific constraints and separate diffusion processes for atom types, atomic positions, and lattice parameters, adding complexity and limiting efficiency. Here, we present DiffCrysGen, a fully data-driven, score-based diffusion model that generates complete crystal structures in a single, end-to-end diffusion process. This unified framework simplifies the model architecture and accelerates sampling by two to three orders of magnitude compared to existing methods without compromising chemical and structural diversity of the generated materials. In order to demonstrate the efficacy of DiffCrysGen in generating valid and useful materials, using density functional theory (DFT), we validate a number of newly generated rare earth-free magnetic materials that are energetically and dynamically stable, and are potentially synthesizable. These include ferromagnets with high saturation magnetization and large magnetocrystalline anisotropy, as also metallic antiferromagnets. These results establish DiffCrysGen as a general platform for accelerated design of functional materials.

Organic surface ligands are integral components of nanocrystals and nanoparticles that have a strong influence on their physicochemical properties, their interaction with the environment, and their ability to self-assemble and order into higher-order structures. These hybrid nanomaterials are tunable with applications in catalysis, directed self-assembly, next-generation optoelectronics, and chemical and quantum sensing. Critically, future advances depend on our ability to rationally engineer their surface chemistry. However, fundamental knowledge of ligand-nanoparticle behavior is limited by uncertainty in where and how these ligands bind to surfaces. For nanoparticles, in particular, few characterization techniques offer both the high spatial resolution and the precise chemical mapping needed to identify specific ligand binding sites. In this study, we utilized synchrotron infrared nanospectroscopy (SINS), atomic force microscopy (AFM), and scanning tunneling microscopy (STM) together with first-principles computer simulations to validate the site-selective adsorption of organic ligands on a shaped nanocrystal surface. Specifically, we demonstrate that the sterically encumbered isocyanide ligands (CNAr^{Mes2}) preferentially bind to the high curvature features of Ag nanocubes (NCs), where low-coordinate Ag atoms are present. In contrast, isocyanide ligands that do not exhibit these steric properties show no surface selectivity. SINS serves as an effective tool to validate these surface binding interactions at the near-single molecule level. These results indicate that steric effects can be successfully harnessed to design bespoke organic ligands for fine-tuning nanocrystal surface chemistry and the properties of the nanocrystal ligand shell.

We introduce ToPolyAgent, a multi-agent AI framework for performing coarse-grained molecular dynamics (MD) simulations of topological polymers through natural language instructions. By integrating large language models (LLMs) with domain-specific computational tools, ToPolyAgent supports both interactive and autonomous simulation workflows across diverse polymer architectures, including linear, ring, brush, and star polymers, as well as dendrimers. The system consists of four LLM-powered agents: a Config Agent for generating initial polymer-solvent configurations, a Simulation Agent for executing LAMMPS-based MD simulations and conformational analyses, a Report Agent for compiling markdown reports, and a Workflow Agent for streamlined autonomous operations. Interactive mode incorporates user feedback loops for iterative refinements, while autonomous mode enables end-to-end task execution from detailed prompts. We demonstrate ToPolyAgent's versatility through case studies involving diverse polymer architectures under varying solvent condition, thermostats, and simulation lengths. Furthermore, we highlight its potential as a research assistant by directing it to investigate the effect of interaction parameters on the linear polymer conformation, and the influence of grafting density on the persistence length of the brush polymer. By coupling natural language interfaces with rigorous simulation tools, ToPolyAgent lowers barriers to complex computational workflows and advances AI-driven materials discovery in polymer science. It lays the foundation for autonomous and extensible multi-agent scientific research ecosystems.

The stability and performance of halide perovskite photovoltaic devices are critically limited by progressive defect generation and associated local non-radiative losses during operation. Self-healing of defects provides a promising pathway to prolong device functionality, yet the underlying microscopic mechanisms remain poorly understood, particularly the role of interfacial chemistry on trap dynamics and healing kinetics. Here, we elucidate self-healing and defect evolution in triple-cation mixed halide (TCMH) perovskite films and their device-relevant charge transport layer heterostructures subjected to photo-induced damage. Using correlation clustering imaging (CLIM), our recently developed local functional imaging tool, we map spatiotemporal photoluminescence heterogeneity to track defect dynamics in pristine and heterostructure films. The defect healing follows bi-phasic kinetics, with an initial electronic relaxation (tens of minutes) and a subsequent slower phase (~ hours) associated to ionic and lattice rearrangement. Most importantly, our results demonstrate that the chemical nature of charge-transport layers modulates trap activity, healing kinetics, and halide redistribution, with heterostructures exhibiting faster recovery than pristine films, a boon for device resilience. These findings provide new insights into the dynamic interaction between defects, interfaces, and ion migration, and establish a framework for rational design of durable, next-generation perovskite optoelectronic devices.

Multiferroic materials, in which electric polarization and magnetic order coexist and couple, offer rich opportunities for both fundamental discovery and technology. However, multiferroicity remains rare due to conflicting electronic requirements for ferroelectricity and magnetism. One route to circumvent this challenge is to exploit the noncollinear ordering of spin cycloids, whose symmetry permits the emergence of polar order. In this work, we introduce another pathway to multiferroic order in which strain generates polarization in materials that host nonpolar spin spirals. To demonstrate this phenomenon, we chose the spin spiral in the well-studied helimagnet Cr1/3NbS2. To detect the induced polarization, we introduce the technique of magnetoelectric birefringence (MEB), an optical probe that enables spatially-resolved and unambiguous detection of polar order. By combining MEB imaging with strain engineering, we confirm the onset of a polar vector at the magnetic transition, establishing strained Cr1/3NbS2 as a type-II multiferroic.

Accurate yet transferable machine-learning interatomic potentials (MLIPs) are essential for accelerating materials and chemical discovery. However, most universal MLIPs overfit to narrow datasets or computational protocols, limiting their reliability across chemical and functional domains. We introduce a transferable multi-domain training strategy that jointly optimizes universal and task-specific parameters through selective regularization, coupled with a domain-bridging set (DBS) that aligns potential-energy surfaces across datasets. Systematic ablation experiments show that small DBS fractions (0.1%) and targeted regularization synergistically enhance out-of-distribution generalization while preserving in-domain fidelity. Trained on fifteen open databases spanning molecules, crystals, and surfaces, our model, SevenNet-Omni, achieves state-of-the-art cross-domain accuracy, including adsorption-energy errors below 0.06 eV on metallic surfaces and 0.1 eV on metal-organic frameworks. Despite containing only 0.5% r$^2$SCAN data, SevenNet-Omni reproduces high-fidelity r$^2$SCAN energetics, demonstrating effective cross-functional transfer from large PBE datasets. This framework offers a scalable route toward universal, transferable MLIPs that bridge quantum-mechanical fidelities and chemical domains.

Relaxor ferroelectric thin films are recognized for their ultrahigh power density, rendering them highly promising for energy storage applications in electrical and electronic systems. However, achieving high energy storage performance with chemically homogeneous, environmentally friendly and compositionally stable materials remains challenging. In this work, we present a design of dielectrics with high energy storage performance via an in-plane polar domains incorporating polar nanoregions mechanism. Guided by phase-field simulations, we synthesized La/Si co-doping BaTiO3 solid-solution thin films with high chemical homogeneity to realize high energy storage performance. Given that, we achieve a high energy density of 203.7J/cm3 and an energy efficiency of approximately 80% at an electric field of 6.15MV/cm. This mechanism holds significant promise for the design of next-generation high-performance dielectric materials for energy storage and other advanced functional materials.

The importance of electrical contact resistance and thermal boundary resistance has increased dramatically as devices are scaled to atomic limits. The use of a rich range of materials with various bandstructures (e.g. parabolic, conical), and in geometries exploiting various dimensionalities (e.g. 1D wires, 2D sheets, and 3D bulk) will increase in the future. Here we derive a single general expression for the quantum limit of electrical contact resistivity for various bandstructures and all dimensions. A corresponding result for the quantum limit of thermal boundary resistance is also derived. These results will be useful to quantitatively co-design, benchmark, and guide the lowering of electrical and thermal boundary resistances for energy efficient devices.

The discovery of flat-bands in magic-angle twisted bilayer graphene has underscored the potential of moire engineering for correlated states, but such phases are notoriously difficult to realize and highly fragile against perturbations. Here, we propose an alternative route to flat-bands by introducing sp3 hybridization in twisted graphite. Instead of relying on fine-tuned magic angles, our approach identifies flat-band states at relatively large twist angles with short moire periods. In this regime, sp3-induced reconstructions generate electronic states that, once formed, are locked by substantial energy barriers, rendering them robust against external perturbations. Using twisted graphite as a prototype, we uncover a series moire-diamond that host two-dimensional flat conduction of valence bands, where carriers are localized within specific momentum planes but remain dispersive along orthogonal directions. The emergence of dimensional flat-bands opens a new platform for flat-band-driven correlated physics and suggests opportunities for designing quantum materials with highly directional electronic functionalities.

Graph generation is a fundamental problem in graph learning with broad applications across Web-scale systems, knowledge graphs, and scientific domains such as drug and material discovery. Recent approaches leverage diffusion models for step-by-step generation, yet unconditional diffusion offers little control over desired properties, often leading to unstable quality and difficulty in incorporating new objectives. Inference-time guidance methods mitigate these issues by adjusting the sampling process without retraining, but they remain inherently local, heuristic, and limited in controllability. To overcome these limitations, we propose TreeDiff, a Monte Carlo Tree Search (MCTS) guided dual-space diffusion framework for controllable graph generation. TreeDiff is a plug-and-play inference-time method that expands the search space while keeping computation tractable. Specifically, TreeDiff introduces three key designs to make it practical and scalable: (1) a macro-step expansion strategy that groups multiple denoising updates into a single transition, reducing tree depth and enabling long-horizon exploration; (2) a dual-space denoising mechanism that couples efficient latent-space denoising with lightweight discrete correction in graph space, ensuring both scalability and structural fidelity; and (3) a dual-space verifier that predicts long-term rewards from partially denoised graphs, enabling early value estimation and removing the need for full rollouts. Extensive experiments on 2D and 3D molecular generation benchmarks, under both unconditional and conditional settings, demonstrate that TreeDiff achieves state-of-the-art performance. Notably, TreeDiff exhibits favorable inference-time scaling: it continues to improve with additional computation, while existing inference-time methods plateau early under limited resources.

Graph generation is a fundamental problem in graph learning with broad applications across Web-scale systems, knowledge graphs, and scientific domains such as drug and material discovery. Recent approaches leverage diffusion models for step-by-step generation, yet unconditional diffusion offers little control over desired properties, often leading to unstable quality and difficulty in incorporating new objectives. Inference-time guidance methods mitigate these issues by adjusting the sampling process without retraining, but they remain inherently local, heuristic, and limited in controllability. To overcome these limitations, we propose TreeDiff, a Monte Carlo Tree Search (MCTS) guided dual-space diffusion framework for controllable graph generation. TreeDiff is a plug-and-play inference-time method that expands the search space while keeping computation tractable. Specifically, TreeDiff introduces three key designs to make it practical and scalable: (1) a macro-step expansion strategy that groups multiple denoising updates into a single transition, reducing tree depth and enabling long-horizon exploration; (2) a dual-space denoising mechanism that couples efficient latent-space denoising with lightweight discrete correction in graph space, ensuring both scalability and structural fidelity; and (3) a dual-space verifier that predicts long-term rewards from partially denoised graphs, enabling early value estimation and removing the need for full rollouts. Extensive experiments on 2D and 3D molecular generation benchmarks, under both unconditional and conditional settings, demonstrate that TreeDiff achieves state-of-the-art performance. Notably, TreeDiff exhibits favorable inference-time scaling: it continues to improve with additional computation, while existing inference-time methods plateau early under limited resources.

Designing protein sequences that fold into a target three-dimensional structure, known as the inverse folding problem, is central to protein engineering but remains challenging due to the vast sequence space and the importance of local structural constraints. Existing deep learning approaches achieve strong recovery rates, yet they lack explicit mechanisms to reuse fine-grained structure-sequence patterns that are conserved across natural proteins. We present PRISM, a multimodal retrieval-augmented generation framework for inverse folding that retrieves fine-grained representations of potential motifs from known proteins and integrates them with a hybrid self-cross attention decoder. PRISM is formulated as a latent-variable probabilistic model and implemented with an efficient approximation, combining theoretical grounding with practical scalability. Across five benchmarks (CATH-4.2, TS50, TS500, CAMEO 2022, and the PDB date split), PRISM establishes new state of the art in both perplexity and amino acid recovery, while also improving foldability metrics (RMSD, TM-score, pLDDT), demonstrating that fine-grained multimodal retrieval is a powerful and efficient paradigm for protein sequence design.

Molecular property prediction is crucial for drug discovery and materials science, yet existing approaches suffer from limited interpretability, poor cross-task generalization, and lack of chemical reasoning capabilities. Traditional machine learning models struggle with task transferability, while specialized molecular language models provide little insight into their decision-making processes. To address these limitations, we propose \textbf{MPPReasoner}, a multimodal large language model that incorporates chemical reasoning for molecular property prediction. Our approach, built upon Qwen2.5-VL-7B-Instruct, integrates molecular images with SMILES strings to enable comprehensive molecular understanding. We develop a two-stage training strategy: supervised fine-tuning (SFT) using 16,000 high-quality reasoning trajectories generated through expert knowledge and multiple teacher models, followed by Reinforcement Learning from Principle-Guided Rewards (RLPGR). RLPGR employs verifiable, rule-based rewards that systematically evaluate chemical principle application, molecular structure analysis, and logical consistency through computational verification. Extensive experiments across 8 datasets demonstrate significant performance improvements, with MPPReasoner outperforming the best baselines by 7.91\% and 4.53\% on in-distribution and out-of-distribution tasks respectively. MPPReasoner exhibits exceptional cross-task generalization and generates chemically sound reasoning paths that provide valuable insights into molecular property analysis, substantially enhancing both interpretability and practical utility for chemists. Code is available at https://anonymous.4open.science/r/MPPReasoner-12687.

Soil moisture (SM) plays a critical role in hydrological and meteorological processes. High-resolution SM can be obtained by combining coarse passive microwave data with fine-scale auxiliary variables. However, the inversion of SM at the temporal scale is hindered by the incompleteness of surface auxiliary factors. To address this issue, first, we introduce validated high temporal resolution ERA5-Land variables into the downscaling process of the low-resolution SMAP SM product. Subsequently, we design a progressive scale convolutional network (PSCNet), at the core of which are two innovative components: a multi-frequency temporal fusion module (MFTF) for capturing temporal dynamics, and a bespoke squeeze-and-excitation (SE) block designed to preserve fine-grained spatial details. Using this approach, we obtained seamless SM products for the Tibetan Plateau (TP) from 2016 to 2018 at 10-km spatial and 3-hour temporal resolution. The experimental results on the TP demonstrated the following: 1) In the satellite product validation, the PSCNet exhibited comparable accuracy and lower error, with a mean R value of 0.881, outperforming other methods. 2) In the in-situ site validation, PSCNet consistently ranked among the top three models for the R metric across all sites, while also showing superior performance in overall error reduction. 3) In the temporal generalization validation, the feasibility of using high-temporal resolution ERA5-Land variables for downscaling was confirmed, as all methods maintained an average relative error within 6\% for the R metric and 2\% for the ubRMSE metric. 4) In the temporal dynamics and visualization validation, PSCNet demonstrated excellent temporal sensitivity and vivid spatial details. Overall, PSCNet provides a promising solution for spatio-temporal downscaling by effectively modeling the intricate spatio-temporal relationships in SM data.

The relevance of this research lies in the growing demand for unmanned aerial vehicles (UAVs) capable of operating reliably in complex environments where conventional navigation becomes unreliable due to interference, poor visibility, or camouflage. Hyperspectral imaging (HSI) provides unique opportunities for UAV-based computer vision by enabling fine-grained material recognition and object differentiation, which are critical for navigation, surveillance, agriculture, and environmental monitoring. The aim of this work is to develop a deep learning architecture integrating HSI into UAV perception for navigation, object detection, and terrain classification. Objectives include: reviewing existing HSI methods, designing a hybrid 2D/3D convolutional architecture with spectral-spatial cross-attention, training, and benchmarking. The methodology is based on the modification of the Mobile 3D Vision Transformer (MDvT) by introducing the proposed SpectralCA block. This block employs bi-directional cross-attention to fuse spectral and spatial features, enhancing accuracy while reducing parameters and inference time. Experimental evaluation was conducted on the WHU-Hi-HongHu dataset, with results assessed using Overall Accuracy, Average Accuracy, and the Kappa coefficient. The findings confirm that the proposed architecture improves UAV perception efficiency, enabling real-time operation for navigation, object recognition, and environmental monitoring tasks. Keywords: SpectralCA, deep learning, computer vision, hyperspectral imaging, unmanned aerial vehicle, object detection, semi-supervised learning.

We propose a generative multivariate posterior sampler via flow matching. It offers a simple training objective, and does not require access to likelihood evaluation. The method learns a dynamic, block-triangular velocity field in the joint space of data and parameters, which results in a deterministic transport map from a source distribution to the desired posterior. The inverse map, named vector rank, is accessible by reversibly integrating the velocity over time. It is advantageous to leverage the dynamic design: proper constraints on the velocity yield a monotone map, which leads to a conditional Brenier map, enabling a fast and simultaneous generation of Bayesian credible sets whose contours correspond to level sets of Monge-Kantorovich data depth. Our approach is computationally lighter compared to GAN-based and diffusion-based counterparts, and is capable of capturing complex posterior structures. Finally, frequentist theoretical guarantee on the consistency of the recovered posterior distribution, and of the corresponding Bayesian credible sets, is provided.

While diffusion models are attracting increasing attention for the design of novel materials, their ability to generate low-energy structures in unexplored chemical spaces has not been systematically assessed. Here, we evaluate the performance of two diffusion models, MatterGen and DiffCSP, against three databases: a ternary oxide set (constructed by a genetic algorithm), a ternary nitride set (constructed by template informatics), and the GNoME database (constructed by a combination of both). We find that diffusion models generally perform stably in well-sampled chemical spaces (oxides and nitrides), but are less effective in uncommon ones (GNoME), which contains many compositions involving rare-earth elements and unconventional stoichiometry. Finally, we assess their size-extrapolation capability and observe a significant drop in performance when the number of atoms exceeds the trained range. This is attributed to the limitations imposed by periodic boundary conditions, which we refer to as the curse of periodicity. This study paves the way for future developments in materials design by highlighting both the strength and the limitations of diffusion models.

While diffusion models are attracting increasing attention for the design of novel materials, their ability to generate low-energy structures in unexplored chemical spaces has not been systematically assessed. Here, we evaluate the performance of two diffusion models, MatterGen and DiffCSP, against three databases: a ternary oxide set (constructed by a genetic algorithm), a ternary nitride set (constructed by template informatics), and the GNoME database (constructed by a combination of both). We find that diffusion models generally perform stably in well-sampled chemical spaces (oxides and nitrides), but are less effective in uncommon ones (GNoME), which contains many compositions involving rare-earth elements and unconventional stoichiometry. Finally, we assess their size-extrapolation capability and observe a significant drop in performance when the number of atoms exceeds the trained range. This is attributed to the limitations imposed by periodic boundary conditions, which we refer to as the curse of periodicity. This study paves the way for future developments in materials design by highlighting both the strength and the limitations of diffusion models.

While diffusion models are attracting increasing attention for the design of novel materials, their ability to generate low-energy structures in unexplored chemical spaces has not been systematically assessed. Here, we evaluate the performance of two diffusion models, MatterGen and DiffCSP, against three databases: a ternary oxide set (constructed by a genetic algorithm), a ternary nitride set (constructed by template informatics), and the GNoME database (constructed by a combination of both). We find that diffusion models generally perform stably in well-sampled chemical spaces (oxides and nitrides), but are less effective in uncommon ones (GNoME), which contains many compositions involving rare-earth elements and unconventional stoichiometry. Finally, we assess their size-extrapolation capability and observe a significant drop in performance when the number of atoms exceeds the trained range. This is attributed to the limitations imposed by periodic boundary conditions, which we refer to as the curse of periodicity. This study paves the way for future developments in materials design by highlighting both the strength and the limitations of diffusion models.

De novo ligand design is a fundamental task that seeks to generate protein or molecule candidates that can effectively dock with protein receptors and achieve strong binding affinity entirely from scratch. It holds paramount significance for a wide spectrum of biomedical applications. However, most existing studies are constrained by the \textbf{Pseudo De Novo}, \textbf{Limited Docking Modeling}, and \textbf{Inflexible Ligand Type}. To address these issues, we propose MagicDock, a forward-looking framework grounded in the progressive pipeline and differentiable surface modeling. (1) We adopt a well-designed gradient inversion framework. To begin with, general docking knowledge of receptors and ligands is incorporated into the backbone model. Subsequently, the docking knowledge is instantiated as reverse gradient flows by binding prediction, which iteratively guide the de novo generation of ligands. (2) We emphasize differentiable surface modeling in the docking process, leveraging learnable 3D point-cloud representations to precisely capture binding details, thereby ensuring that the generated ligands preserve docking validity through direct and interpretable spatial fingerprints. (3) We introduce customized designs for different ligand types and integrate them into a unified gradient inversion framework with flexible triggers, thereby ensuring broad applicability. Moreover, we provide rigorous theoretical guarantees for each component of MagicDock. Extensive experiments across 9 scenarios demonstrate that MagicDock achieves average improvements of 27.1\% and 11.7\% over SOTA baselines specialized for protein or molecule ligand design, respectively.

The stochastic multi-armed bandit (MAB) problem is one of the most fundamental models in sequential decision-making, with the core challenge being the trade-off between exploration and exploitation. Although algorithms such as Upper Confidence Bound (UCB) and Thompson Sampling, along with their regret theories, are well-established, existing analyses primarily operate from a time-domain and cumulative regret perspective, struggling to characterize the dynamic nature of the learning process. This paper proposes a novel frequency-domain analysis framework, reformulating the bandit process as a signal processing problem. Within this framework, the reward estimate of each arm is viewed as a spectral component, with its uncertainty corresponding to the component's frequency, and the bandit algorithm is interpreted as an adaptive filter. We construct a formal Frequency-Domain Bandit Model and prove the main theorem: the confidence bound term in the UCB algorithm is equivalent in the frequency domain to a time-varying gain applied to uncertain spectral components, a gain inversely proportional to the square root of the visit count. Based on this, we further derive finite-time dynamic bounds concerning the exploration rate decay. This theory not only provides a novel and intuitive physical interpretation for classical algorithms but also lays a rigorous theoretical foundation for designing next-generation algorithms with adaptive parameter adjustment.

In-context learning allows large models to adapt to new tasks from a few demonstrations, but it has shown limited success in molecular design. Existing databases such as ChEMBL contain molecular properties spanning millions of biological assays, yet labeled data for each property remain scarce. To address this limitation, we introduce demonstration-conditioned diffusion models (DemoDiff), which define task contexts using a small set of molecule-score examples instead of text descriptions. These demonstrations guide a denoising Transformer to generate molecules aligned with target properties. For scalable pretraining, we develop a new molecular tokenizer with Node Pair Encoding that represents molecules at the motif level, requiring 5.5$\times$ fewer nodes. We curate a dataset containing millions of context tasks from multiple sources covering both drugs and materials, and pretrain a 0.7-billion-parameter model on it. Across 33 design tasks in six categories, DemoDiff matches or surpasses language models 100-1000$\times$ larger and achieves an average rank of 3.63 compared to 5.25-10.20 for domain-specific approaches. These results position DemoDiff as a molecular foundation model for in-context molecular design. Our code is available at https://github.com/liugangcode/DemoDiff.

Energetic electrons and holes generated from the decay of localized surface plasmons in metallic nanoparticles can be harnessed in nanoscale devices for photocatalysis, photovoltaics or sensing. In this work, we study the generation of such hot carriers in bimetallic Janus nanoparticles composed of Au, Ag and Cu using a recently developed atomistic modelling approach that combines a solution of the macroscopic Maxwell equation with large-scale quantum-mechanical tight-binding models. We first analyze spherical Janus nanoparticles whose unique hot-carrier spectrum can be associated with the spectra of the two hemispheres and the interface coupling and find that under solar illumination the Ag-Au system exhibits the highest hot-carrier generation rate. For dumbbell-shaped Janus nanoparticles, we observe a significant increase in hot-carrier generation with increasing neck size. This is caused by a dramatic enhancement of the electric field in the neck region. We also study the dependence of hot-carrier generation on the light polarization and find that the largest generation rates are obtained when the electric field is perpendicular to the interface between the two metals due to the maximal dipole coupling with the electric field. The insights from our study will guide the experimental design of efficient hot-carrier devices based on bimetallic Janus nanoparticles.

Recent developments have enabled advanced reasoning in Large Language Models (LLMs) via long Chain-of-Thought (CoT), while long CoT suffers from high computational costs and significant latency losses owing to the autoregressive nature of generative LLMs. CoT compression aims to improve efficiency in the reasoning process by reducing output length. Previous works trade reasoning efficiency by either laborious discrete prompt designing or the construction of external compressed CoT datasets that sacrifice key reasoning details. In this work, we propose Upfront CoT (UCoT): an efficient reasoning framework with upfront thought embedding to automate CoT compression. UCoT is a cooperative workflow involving a small model (compressor) and a large model (executor). The first stage of UCoT trains compressor to generate upfront thought embeddings rich in reasoning information for the executor, avoiding the drawbacks of manually designed prompts. The second stage optimizes executor to utilize upfront thought embeddings to derive the correct answer with short reasoning, using a reward mechanism. Extensive experiments show that UCoT maintains the powerful reasoning ability of executor while significantly reducing the length of CoT. It is worth mentioning that when applying UCoT to the Qwen2.5-7B-Instruct model, the usage of tokens on GSM8K dataset is reduced by 50\%, while the performance is 3.08\% higher than that of the state-of-the-art (SOTA) method. The code and dataset are in supplementary material.

Molecular representation learning plays a crucial role in advancing applications such as drug discovery and material design. Existing work leverages 2D and 3D modalities of molecular information for pre-training, aiming to capture comprehensive structural and geometric insights. However, these methods require paired 2D and 3D molecular data to train the model effectively and prevent it from collapsing into a single modality, posing limitations in scenarios where a certain modality is unavailable or computationally expensive to generate. To overcome this limitation, we propose FlexMol, a flexible molecule pre-training framework that learns unified molecular representations while supporting single-modality input. Specifically, inspired by the unified structure in vision-language models, our approach employs separate models for 2D and 3D molecular data, leverages parameter sharing to improve computational efficiency, and utilizes a decoder to generate features for the missing modality. This enables a multistage continuous learning process where both modalities contribute collaboratively during training, while ensuring robustness when only one modality is available during inference. Extensive experiments demonstrate that FlexMol achieves superior performance across a wide range of molecular property prediction tasks, and we also empirically demonstrate its effectiveness with incomplete data. Our code and data are available at https://github.com/tewiSong/FlexMol.

Moiré materials offer a versatile platform for engineering excitons with unprecedented control, promising next-generation optoelectronic applications. While continuum models are widely used to study moiré excitons due to their computational efficiency, they often disagree with ab initio many-body approaches, as seen for intralayer excitons in WS$_2$/WSe$_2$ heterobilayers. Here, we resolve these discrepancies using an atomistic, quantum-mechanical framework based on the Bethe-Salpeter equation with localized Wannier functions as the basis for the electronic structure. We show that inclusion of dielectric screening due to hexagonal boron nitride (hBN) encapsulation is essential to reproduce the full set of experimentally observed features of moiré intralayer excitons. Our analysis reveals a competition between Wannier and charge transfer characters, driven by variations between direct and indirect band gaps at high symmetry stacking regions due to atomic relaxations and environmentally tunable electron-hole interactions. Building on this insight, we demonstrate that the lowest-energy bright excitons are Wannier-like in WS2/WSe2 heterobilayers but charge-transfer-like in twisted WSe2 homobilayers, despite having comparable moiré lengths when encapsulated in hBN. In the absence of hBN encapsulation, the lowest-energy bright exciton in twisted WSe$_2$ becomes Wannier-like. These results establish atomistic modeling as a powerful and efficient approach for designing and controlling excitonic phenomena in moiré materials.

AI models are increasingly used for data analysis and visualization, yet benchmarks rarely address scatterplot-specific tasks, limiting insight into performance. To address this gap for one of the most common chart types, we introduce a synthetic, annotated dataset of over 18,000 scatterplots from six data generators and 17 chart designs, and a benchmark based on it. We evaluate proprietary models from OpenAI and Google using N-shot prompting on five distinct tasks derived from annotations of cluster bounding boxes, their center coordinates, and outlier coordinates. OpenAI models and Gemini 2.5 Flash, especially when prompted with examples, are viable options for counting clusters and, in Flash's case, outliers (90%+ Accuracy). However, the results for localization-related tasks are unsatisfactory: Precision and Recall are near or below 50%, except for Flash in outlier identification (65.01%). Furthermore, the impact of chart design on performance appears to be a secondary factor, but it is advisable to avoid scatterplots with wide aspect ratios (16:9 and 21:9) or those colored randomly. Supplementary materials are available at https://github.com/feedzai/biy-paper.

Integrating monolayers derived from 2D van der Waals (vdW) magnetic materials into next-generation technological applications remains a significant challenge due to their structural and magnetic instability issues. Template-assisted encapsulation is a potential route for the growth of stable 2D monolayers aimed at designing novel 1D heterostructures, opening new avenues for studying low-dimensional quantum effects and spin-related phenomena. In this study, we explored the diameter-dependent encapsulation of 2D CrI3 crystals using multi-walled carbon nanotubes as nanoscale host templates. Advanced microscopic analysis revealed distinct structural transitions, ranging from internal nanorod encapsulation to external shell formation, directly influenced by the host nanotube diameter. Furthermore, statistical analysis of structural morphologies indicates that CrI3 nanorods preferentially form within MWCNTs with inner diameters up to 5 nm, while single-walled CrI3 nanotubes are stabilized in CNTs with diameters up to 8 nm. For host CNTs exceeding 10 nm in diameter, CrI3 predominantly forms surface coatings rather than confined one-dimensional structures. In situ electron beam irradiation demonstrates the superior structural stability of single-walled CrI3 confined within MWCNTs, while externally coated CrI3 undergoes decomposition into metallic Cr clusters. Prolonged irradiation induces a morphological transformation of CrI3 nanotubes into nanorods. These insights lay the groundwork for engineering robust, tunable 1D magnetic heterostructures of CrI3 for spintronic and data storage applications.

Accurate simulation of wave propagation in complex acoustic materials is crucial for applications in sound design, noise control, and material engineering. Traditional numerical solvers, such as finite element methods, are computationally expensive, especially when dealing with large-scale or real-time scenarios. In this work, we introduce a dataset of 31,000 acoustic materials, named HA30K, designed and simulated solving the Helmholtz equations. For each material, we provide the geometric configuration and the corresponding pressure field solution, enabling data-driven approaches to learn Helmholtz equation solutions. As a baseline, we explore a deep learning approach based on Stable Diffusion with ControlNet, a state-of-the-art model for image generation. Unlike classical solvers, our approach leverages GPU parallelization to process multiple simulations simultaneously, drastically reducing computation time. By representing solutions as images, we bypass the need for complex simulation software and explicit equation-solving. Additionally, the number of diffusion steps can be adjusted at inference time, balancing speed and quality. We aim to demonstrate that deep learning-based methods are particularly useful in early-stage research, where rapid exploration is more critical than absolute accuracy.

Guiding pretrained flow-based generative models for conditional generation or to produce samples with desired target properties enables solving diverse tasks without retraining on paired data. We present ESS-Flow, a gradient-free method that leverages the typically Gaussian prior of the source distribution in flow-based models to perform Bayesian inference directly in the source space using Elliptical Slice Sampling. ESS-Flow only requires forward passes through the generative model and observation process, no gradient or Jacobian computations, and is applicable even when gradients are unreliable or unavailable, such as with simulation-based observations or quantization in the generation or observation process. We demonstrate its effectiveness on designing materials with desired target properties and predicting protein structures from sparse inter-residue distance measurements.

Molecular dynamics (MD) simulations underpin modern computational drug dis- covery, materials science, and biochemistry. Recent machine learning models provide high-fidelity MD predictions without the need to repeatedly solve quantum mechanical forces, enabling significant speedups over conventional pipelines. Yet many such methods typically enforce strict equivariance and rely on sequential rollouts, thus limiting their flexibility and simulation efficiency. They are also com- monly single-task, trained on individual molecules and fixed timeframes, which restricts generalization to unseen compounds and extended timesteps. To address these issues, we propose Atomistic Transformer Operator for Molecules (ATOM), a pretrained transformer neural operator for multitask molecular dynamics. ATOM adopts a quasi-equivariant design that requires no explicit molecular graph and employs a temporal attention mechanism, allowing for the accurate parallel decod- ing of multiple future states. To support operator pretraining across chemicals and timescales, we curate TG80, a large, diverse, and numerically stable MD dataset with over 2.5 million femtoseconds of trajectories across 80 compounds. ATOM achieves state-of-the-art performance on established single-task benchmarks, such as MD17, RMD17 and MD22. After multitask pretraining on TG80, ATOM shows exceptional zero-shot generalization to unseen molecules across varying time hori- zons. We believe ATOM represents a significant step toward accurate, efficient, and transferable molecular dynamics models

Accurate and efficient temperature prediction is critical for optimizing the preheating process of PET preforms in industrial microwave systems prior to blow molding. We propose a novel deep learning framework for generalized temperature prediction. Unlike traditional models that require extensive retraining for each material or design variation, our method introduces a data-efficient neural architecture that leverages transfer learning and model fusion to generalize across unseen scenarios. By pretraining specialized neural regressor on distinct conditions such as recycled PET heat capacities or varying preform geometries and integrating their representations into a unified global model, we create a system capable of learning shared thermal dynamics across heterogeneous inputs. The architecture incorporates skip connections to enhance stability and prediction accuracy. Our approach reduces the need for large simulation datasets while achieving superior performance compared to models trained from scratch. Experimental validation on two case studies material variability and geometric diversity demonstrates significant improvements in generalization, establishing a scalable ML-based solution for intelligent thermal control in manufacturing environments. Moreover, the approach highlights how data-efficient generalization strategies can extend to other industrial applications involving complex physical modeling with limited data.

Solar thermochemical water splitting enables hydrogen production by cycling metal oxides between reduced and oxidized states, typically through an oxygen vacancy mechanism. However, recent experimental work suggests that cation vacancies have a greater influence on the redox behavior of iron aluminate spinels used in water splitting. This remains debated, as calculations predict that such cation vacancies are thermodynamically unfavorable. In the current work, we show that Fe vacancies in (Fe$ζ$Al1-$ζ$)3O4 become accessible only when facilitated by inversion between Fe and Al. This antisite disorder lowers the formation energy of octahedral Fe vacancies in Al-rich spinels ($ζ$ = 1/3) from over 3 eV to just 0.62 eV when one third of the cation sites are inverted, allowing high Fe vacancy concentrations under oxidizing conditions. This mechanism supports high H2 yields up to 361 $μ$mol/g, consistent with experimental observations. Our findings support the notion that solar thermochemical water splitting can occur through a cation vacancy mechanism. They also clarify how site inversion, vacancy energetics, and defect interactions each contribute to redox performance, offering general design principles for identifying and optimizing materials that operate through cation vacancy cycling.

Atomistic machine learning (ML) is a transformative tool for accurate and efficient investigation of material behavior at the atomic scale. While such models have been constructed within Cartesian space to harness geometric information and preserve intuitive physical representations, they face inherent challenges - primarily due to the lack of a systematic symmetry-preserving framework for representing arbitrary physical tensors. We address these challenges by proposing Cartesian Natural Tensor Networks (CarNet) as a general framework for atomistic ML. We first develop the theory of irreducible representations using Cartesian natural tensors (their creation, operation, as well as the decomposition and reconstruction of physical tensors such as the elastic constant tensor). Leveraging this machinery, we design an equivariant Cartesian model and demonstrate its exceptional performance across diverse atomistic ML tasks. CarNet enables the development of highly accurate and reliable interatomic potentials for both materials and molecular systems. Furthermore, structure-property relationships can be readily constructed for tensorial quantities ranging from simple properties like the dipole moment to arbitrary high-rank tensors with complex symmetries such as the elastic constant tensor -- capabilities that were previously inaccessible. This work removes theoretical barriers and unleashes the power of Cartesian approaches for advanced atomistic ML in the understanding and design of new materials.

Small language models (SLMs; 1-12B params, sometimes up to 20B) are sufficient and often superior for agentic workloads where the objective is schema- and API-constrained accuracy rather than open-ended generation. We synthesize recent evidence across open and proprietary SLMs (Phi-4-Mini, Qwen-2.5-7B, Gemma-2-9B, Llama-3.2-1B/3B, Ministral-3B/8B, Apple on-device 3B, DeepSeek-R1-Distill) and connect it to modern evaluations (BFCL v3/v4, StableToolBench) and serving stacks (vLLM, SGLang, TensorRT-LLM) paired with guided decoding libraries (XGrammar, Outlines). We formalize SLM-default, LLM-fallback systems with uncertainty-aware routing and verifier cascades, and propose engineering metrics that reflect real production goals: cost per successful task (CPS), schema validity rate, executable call rate, p50/p95 latency, and energy per request. Guided decoding, strict JSON Schema outputs, and validator-first tool execution close much of the capability gap with larger models and often let SLMs match or surpass LLMs on tool use, function calling, and RAG at 10x-100x lower token cost with materially better latency and energy. We provide design patterns for agent stacks that prioritize SLMs: schema-first prompting, type-safe function registries, confidence scoring with verifier rollups, and lightweight adaptation via LoRA/QLoRA. We also delineate limits where fallback remains valuable (open-domain reasoning and some long-horizon planning). The result is a practical blueprint for building fast, inexpensive, and reliable agents that default to SLMs while preserving headroom with targeted LLM assistance. Keywords: small language models, agents, function calling, structured outputs, JSON Schema, guided decoding, LoRA/QLoRA, routing, energy efficiency, edge inference

The focused ion beam (FIB) microscope is well established as a high-resolution machining instrument capable of site-selectively removing material down to the nanoscale. Beyond subtractive processing, however, the FIB can also add material using a technique known as focused ion beam induced deposition (FIBID), enabling the direct-write of complex nanostructures. This work explores new directions in three-dimensional nanoprinting with FIBID, harnessing unique features of helium and neon FIBs to fabricate nanoscale heterostructures, including multimaterial architectures and deposits with engineered internal voids. Detailed insight into the chemical and structural composition of these nanostructures is obtained using advanced electron microscopy, revealing buried interfaces and material transformations. Building on these results, the evolution of FIBID into a versatile platform for functional nanomaterials design is discussed, opening pathways toward next-generation nanoscale devices and technologies.

High-spin states hold significant promise for classical and quantum information storage and emerging magnetic memory technologies. Here, we present a systematic framework for engineering such high-spin magnetic states in dopant clusters formed from substitutional impurities in semiconductors. In single-valley materials such as gallium arsenide, impurity states are hydrogenic and exchange interactions generally favor low-spin configurations, except in special geometries. In contrast, multivalley semiconductors exhibit oscillatory form factors in their exchange couplings, enabling the controlled suppression of selected hopping processes and exchange couplings. Exploiting this feature, we demonstrate how carefully arranged impurities in aluminum arsenide, germanium, and silicon can stabilize ground states with a net spin that scale extensively with system size. Within effective mass theory and the tight-binding approximation for hopping, we construct explicit examples ranging from finite clusters to extended lattices and fractal-like tilings. In two dimensions, we identify several favorable dopant geometries supporting a net spin equal to around half of the fully polarized value in the thermodynamic limit, including one which achieves over $70\%$ polarization. Our results provide a general design principle for harnessing valley degeneracy in semiconductors to construct robust high-spin states and outline a pathway for their experimental realization via precision implantation of dopants.

Inverse design, which seeks to find optimal parameters for a target output, is a central challenge in engineering. Surrogate-based optimization (SBO) has become a standard approach, yet it is fundamentally structured to converge to a single-point solution, thereby limiting design space exploration and ignoring potentially valuable alternative topologies. This paper presents a paradigm shift from single-point optimization to generative inverse design. We introduce a framework based on a Conditional Variational Autoencoder (CVAE) that learns a probabilistic mapping between a system's design parameters and its performance, enabling the generation of a diverse portfolio of high-performing candidates conditioned on a specific performance objective. We apply this methodology to the complex, non-linear problem of minimizing airfoil self-noise, using a high-performing SBO method from a prior benchmark study as a rigorous baseline. The CVAE framework successfully generated 256 novel designs with a 94.1\% validity rate. A subsequent surrogate-based evaluation revealed that 77.2\% of these valid designs achieved superior performance compared to the single optimal design found by the SBO baseline. This work demonstrates that the generative approach not only discovers higher-quality solutions but also provides a rich portfolio of diverse candidates, fundamentally enhancing the engineering design process by enabling multi-criteria decision-making.

Agentic systems enable the intelligent use of research tooling, augmenting a researcher's ability to investigate and propose novel solutions to existing problems. Within Additive Manufacturing (AM), alloy discovery remains a complex challenge, often requiring expertise in the various domains of materials science, thermodynamic simulations, and experimental analysis. Large Language Model (LLM) enabled agents can facilitate this endeavor by utilizing their extensive knowledge base to dispatch tool calls via Model Context Protocol (MCP) to perform actions such as Thermo-Calc property diagram calculations and lack of fusion process map generation. In addition, the multi-agent system developed in this work is able to effectively reason through complex user prompts and provide analysis on the printability of proposed alloys. These agents can dynamically adjust their task trajectory to the outcomes of tool call results, effectively enabling autonomous decision-making in practical environments. This work aims to utilize LLM enabled agents to automate and accelerate the task of alloy discovery within the field of additive manufacturing and showcase the benefits of adopting this multi-agent system.

Surface oxides are associated with two-level systems (TLSs) that degrade the performance of niobium-based superconducting quantum computing devices. To address this, we introduce a predictive framework for selecting metal capping layers that inhibit niobium oxide formation. Using DFT-calculated oxygen interstitial and vacancy energies as thermodynamic descriptors, we train a logistic regression model on a limited set of experimental outcomes to successfully predict the likelihood of oxide formation beneath different capping materials. This approach identifies Zr, Hf, and Ta as effective diffusion barriers. Our analysis further reveals that the oxide formation energy per oxygen atom serves as an excellent standalone descriptor for predicting barrier performance. By combining this new descriptor with lattice mismatch as a secondary criterion to promote structurally coherent interfaces, we identify Zr, Ta, and Sc as especially promising candidates. This closed-loop strategy integrates first-principles theory, machine learning, and limited experimental data to enable rational design of next-generation materials.

Diffusion models have emerged as powerful generative priors for solving inverse imaging problems. However, their practical deployment is hindered by the substantial computational cost of slow, multi-step sampling. Although Consistency Models (CMs) address this limitation by enabling high-quality generation in only one or a few steps, their direct application to inverse problems has remained largely unexplored. This paper introduces a modified consistency sampling framework specifically designed for inverse problems. The proposed approach regulates the sampler's stochasticity through a measurement-consistency mechanism that leverages the degradation operator, thereby enforcing fidelity to the observed data while preserving the computational efficiency of consistency-based generation. Comprehensive experiments on the Fashion-MNIST and LSUN Bedroom datasets demonstrate consistent improvements across both perceptual and pixel-level metrics, including the Fréchet Inception Distance (FID), Kernel Inception Distance (KID), peak signal-to-noise ratio (PSNR), and structural similarity index measure (SSIM), compared with baseline consistency and diffusion-based sampling methods. The proposed method achieves competitive or superior reconstruction quality with only a small number of sampling steps.

Generalization in deep learning is closely tied to the pursuit of flat minima in the loss landscape, yet classical Stochastic Gradient Langevin Dynamics (SGLD) offers no mechanism to bias its dynamics toward such low-curvature solutions. This work introduces Flatness-Aware Stochastic Gradient Langevin Dynamics (fSGLD), designed to efficiently and provably seek flat minima in high-dimensional nonconvex optimization problems. At each iteration, fSGLD uses the stochastic gradient evaluated at parameters perturbed by isotropic Gaussian noise, commonly referred to as Random Weight Perturbation (RWP), thereby optimizing a randomized-smoothing objective that implicitly captures curvature information. Leveraging these properties, we prove that the invariant measure of fSGLD stays close to a stationary measure concentrated on the global minimizers of a loss function regularized by the Hessian trace whenever the inverse temperature and the scale of random weight perturbation are properly coupled. This result provides a rigorous theoretical explanation for the benefits of random weight perturbation. In particular, we establish non-asymptotic convergence guarantees in Wasserstein distance with the best known rate and derive an excess-risk bound for the Hessian-trace regularized objective. Extensive experiments on noisy-label and large-scale vision tasks, in both training-from-scratch and fine-tuning settings, demonstrate that fSGLD achieves superior or comparable generalization and robustness to baseline algorithms while maintaining the computational cost of SGD, about half that of SAM. Hessian-spectrum analysis further confirms that fSGLD converges to significantly flatter minima.

Efficient and inexpensive energy storage is essential for accelerating the adoption of renewable energy and ensuring a stable supply, despite fluctuations in sources such as wind and solar. Electrocatalysts play a key role in hydrogen energy storage (HES), allowing the energy to be stored as hydrogen. However, the development of affordable and high-performance catalysts for this process remains a significant challenge. We introduce Catalyst GFlowNet, a generative model that leverages machine learning-based predictors of formation and adsorption energy to design crystal surfaces that act as efficient catalysts. We demonstrate the performance of the model through a proof-of-concept application to the hydrogen evolution reaction, a key reaction in HES, for which we successfully identified platinum as the most efficient known catalyst. In future work, we aim to extend this approach to the oxygen evolution reaction, where current optimal catalysts are expensive metal oxides, and open the search space to discover new materials. This generative modeling framework offers a promising pathway for accelerating the search for novel and efficient catalysts.

Pose estimation refers to tracking a human's full body posture, including their head, torso, arms, and legs. The problem is challenging in practical settings where the number of body sensors are limited. Past work has shown promising results using conditional diffusion models, where the pose prediction is conditioned on both measurements from the sensors. Unfortunately, nearly all these approaches generalize poorly across users, primarly because location measurements are highly influenced by the body size of the user. In this paper, we formulate pose estimation as an inverse problem and design an algorithm capable of zero-shot generalization. Our idea utilizes a pre-trained diffusion model and conditions it on rotational measurements alone; the priors from this model are then guided by a likelihood term, derived from the measured locations. Thus, given any user, our proposed InPose method generatively estimates the highly likely sequence of poses that best explains the sparse on-body measurements.

This work introduces ShapeGen3DCP, a deep learning framework for fast and accurate prediction of filament cross-sectional geometry in 3D Concrete Printing (3DCP). The method is based on a neural network architecture that takes as input both material properties in the fluid state (density, yield stress, plastic viscosity) and process parameters (nozzle diameter, nozzle height, printing and flow velocities) to directly predict extruded layer shapes. To enhance generalization, some inputs are reformulated into dimensionless parameters that capture underlying physical principles. Predicted geometries are compactly represented using Fourier descriptors, which enforce smooth, closed, and symmetric profiles while reducing the prediction task to a small set of coefficients. The training dataset was synthetically generated using a well-established Particle Finite Element (PFEM) model of 3DCP, overcoming the scarcity of experimental data. Validation against diverse numerical and experimental cases shows strong agreement, confirming the framework's accuracy and reliability. This opens the way to practical uses ranging from pre-calibration of print settings, minimizing or even eliminating trial-and-error adjustments, to toolpath optimization for more advanced designs. Looking ahead, coupling the framework with simulations and sensor feedback could enable closed-loop digital twins for 3DCP, driving real-time process optimization, defect detection, and adaptive control of printing parameters.

Multi-layer perceptrons (MLPs) conventionally follow a narrow-wide-narrow design where skip connections operate at the input/output dimensions while processing occurs in expanded hidden spaces. We challenge this convention by proposing wide-narrow-wide (Hourglass) MLP blocks where skip connections operate at expanded dimensions while residual computation flows through narrow bottlenecks. This inversion leverages higher-dimensional spaces for incremental refinement while maintaining computational efficiency through parameter-matched designs. Implementing Hourglass MLPs requires an initial projection to lift input signals to expanded dimensions. We propose that this projection can remain fixed at random initialization throughout training, enabling efficient training and inference implementations. We evaluate both architectures on generative tasks over popular image datasets, characterizing performance-parameter Pareto frontiers through systematic architectural search. Results show that Hourglass architectures consistently achieve superior Pareto frontiers compared to conventional designs. As parameter budgets increase, optimal Hourglass configurations favor deeper networks with wider skip connections and narrower bottlenecks-a scaling pattern distinct from conventional MLPs. Our findings suggest reconsidering skip connection placement in modern architectures, with potential applications extending to Transformers and other residual networks.

Engineered micro- or nano-structures based on nonlinear optical materials offer versatile opportunities for optoelectronic applications. While extensive efforts have been devoted to design tailored microcavities to promote and increase the optical nonlinearities of graphene, the potential of engineering its three-dimensional counterparts -- three-dimensional Dirac semimetals -- remains largely unexplored. Here we report on exceptionally strong terahertz nonlinearities in a cavity-engineered Dirac semimetal microstructure, and demonstrate a giant enhancement of terahertz third- and fifth-order harmonic yields by more than three orders of magnitude. By fabricating a designed structure of metallic metasurface microcavities on a nanometer thin film of the threedimensional Dirac semimetal Cd3As2, we significantly enhance the near-field intensity of a picosecond terahertz excitation pulse in resonance with the microcavity eigenmode. This transiently modifies the nonlinearities of the thin film and drives the nonlinear responses of the Dirac fermions from a weakly to a deeply nonperturbative regime where the observed high-harmonic generation essentially saturates.

Data-driven materials discovery requires large-scale experimental datasets, yet most of the information remains trapped in unstructured literature. Existing extraction efforts often focus on a limited set of features and have not addressed the integrated composition-processing-microstructure-property relationships essential for understanding materials behavior, thereby posing challenges for building comprehensive databases. To address this gap, we propose a multi-stage information extraction pipeline powered by large language models, which captures 47 features spanning composition, processing, microstructure, and properties exclusively from experimentally reported materials. The pipeline integrates iterative extraction with source tracking to enhance both accuracy and reliability. Evaluations at the feature level (independent attributes) and tuple level (interdependent features) yielded F1 scores around 0.96. Compared with single-pass extraction without source tracking, our approach improved F1 scores of microstructure category by 10.0% (feature level) and 13.7% (tuple level), and reduced missed materials from 49 to 13 out of 396 materials in 100 articles on precipitate-containing multi-principal element alloys (miss rate reduced from 12.4% to 3.3%). The pipeline enables scalable and efficient literature mining, producing databases with high precision, minimal omissions, and zero false positives. These datasets provide trustworthy inputs for machine learning and materials informatics, while the modular design generalizes to diverse material classes, enabling comprehensive materials information extraction.

The ability to model mechanics of soft materials under flowing conditions is key in designing and engineering processes and materials with targeted properties. This generally requires solution of internal stress tensor, related to the deformation tensor through nonlinear and history-dependent constitutive models. Traditional numerical methods for non-Newtonian fluid dynamics often suffer from prohibitive computational demands and poor scalability to new problem instances. Developments in data-driven methods have mitigated some limitations but still require retraining across varied physical conditions. In this work, we introduce Rheological Operator Transformer (RheOFormer), a generative operator learning method leveraging self-attention to efficiently learn different spatial interactions and features of complex fluid flows. We benchmark RheOFormer across a range of different viscometric and non-viscometric flows with different types of viscoelastic and elastoviscoplastic mechanics in complex domains against ground truth solutions. Our results demonstrate that RheOFormer can accurately learn both scalar and tensorial nonlinear mechanics of different complex fluids and predict the spatio-temporal evolution of their flows, even when trained on limited datasets. Its strong generalization capabilities and computational efficiency establish RheOFormer as a robust neural surrogate for accelerating predictive complex fluid simulations, advancing data-driven experimentation, and enabling real-time process optimization across a wide range of applications.

We present exa-AMD, an open-source, high-performance framework designed for accelerated materials discovery on modern supercomputers. exa-AMD overcomes key computational bottlenecks in large-scale structure prediction through task-based parallelization, adaptive load balancing, and optimized data management for CPU and GPU architectures. The framework automates the end-to-end workflow, from generating candidate structures to evaluating formation energies and updating phase diagrams. Its modular design allows users to easily replace or extend components with custom machine learning models, alternative initial structure templates, and future structure generators, enabling flexible integration with emerging AI approaches. We demonstrate strong scaling across high-performance computing platforms and highlight applications to Na-B-C, Ce-Co-B, and Fe-Co-Zr systems, establishing exa-AMD as a robust and exascale-ready tool for accelerating the discovery and design of functional materials. exa-AMD is publicly available on GitHub, with detailed documentation and reproducible test cases to support community engagement and collaborative research.

The emergence of agent-based systems represents a significant advancement in artificial intelligence, with growing applications in automated data extraction. However, chemical information extraction remains a formidable challenge due to the inherent heterogeneity of chemical data. Current agent-based approaches, both general-purpose and domain-specific, exhibit limited performance in this domain. To address this gap, we present ChemX, a comprehensive collection of 10 manually curated and domain-expert-validated datasets focusing on nanomaterials and small molecules. These datasets are designed to rigorously evaluate and enhance automated extraction methodologies in chemistry. To demonstrate their utility, we conduct an extensive benchmarking study comparing existing state-of-the-art agentic systems such as ChatGPT Agent and chemical-specific data extraction agents. Additionally, we introduce our own single-agent approach that enables precise control over document preprocessing prior to extraction. We further evaluate the performance of modern baselines, such as GPT-5 and GPT-5 Thinking, to compare their capabilities with agentic approaches. Our empirical findings reveal persistent challenges in chemical information extraction, particularly in processing domain-specific terminology, complex tabular and schematic representations, and context-dependent ambiguities. The ChemX benchmark serves as a critical resource for advancing automated information extraction in chemistry, challenging the generalization capabilities of existing methods, and providing valuable insights into effective evaluation strategies.

In this paper, we present a vocoder-free framework for audio super-resolution that employs a flow matching generative model to capture the conditional distribution of complex-valued spectral coefficients. Unlike conventional two-stage diffusion-based approaches that predict a mel-spectrogram and then rely on a pre-trained neural vocoder to synthesize waveforms, our method directly reconstructs waveforms via the inverse Short-Time Fourier Transform (iSTFT), thereby eliminating the dependence on a separate vocoder. This design not only simplifies end-to-end optimization but also overcomes a critical bottleneck of two-stage pipelines, where the final audio quality is fundamentally constrained by vocoder performance. Experiments show that our model consistently produces high-fidelity 48 kHz audio across diverse upsampling factors, achieving state-of-the-art performance on both speech and general audio datasets.

To simultaneously achieve high hardness and high toughness in protective coatings remains a fundamental challenge. Here, we harness the superlattice architecture to combine Koehler hardening while the coherent interfaces reduce the crack driving force and improve toughness, enabling coatings that are both hard and damage tolerant. We design and fabricate epitaxial HfN$_{1.33}$/Hf$_{0.76}$Al$_{0.24}$N$_{1.15}$ superlattices, deposited on MgO(001) substrates using low-energy, high-flux ion-assisted reactive magnetron sputtering. These superlattices with bilayer periods ranging from 6 to 20 nm, exhibit a unique three-fold superstructure, confirmed by X-ray diffraction and reciprocal space mapping (RSM). Each constituent forms distinct 3D checkerboard superstructures, with a period of 7.5 Å for HfN and 12.5 A for HfAlN. RSMs further reveal low mosaicity, high crystalline quality, and in-plane compressive strains, indicating well preserved coherence across interfaces. Mechanical testing shows that the superlattices maintain the high hardness of HfAlN (\~36 GPa) independent of bilayer period, while surpassing the softer HfN (~27 GPa), consistent with interface-driven Koehler strengthening. Micropillar compression shows brittle fracture on the {110}<110> system, yet with distributed cracking and faster mechanical recovery compared to monolithic films, suggesting improved toughness. Cube-corner indentation further corroborate this behavior, with pile-up and suppressed fracture events. These results demonstrate that epitaxial HfN/HfAlN superlattices uniquely combine high hardness with improved toughness, enabled by their three-fold superstructured architecture. Leveraging the intrinsic high-temperature stability of HfN-based materials, this design offers a robust pathway toward next-generation protective coatings capable of maintaining performance under extreme conditions.

While Retrieval-Augmented Generation (RAG) enables large language models (LLMs) to generate contextually grounded responses, contextual faithfulness remains challenging as LLMs may not consistently trust provided context, leading to hallucinations that undermine reliability. We observe an inverse correlation between response copying degree and context-unfaithful hallucinations on RAGTruth, suggesting that higher copying degrees reduce hallucinations by fostering genuine contextual belief. We propose CopyPasteLLM, obtained through two-stage high-copying response preference training. We design three prompting methods to enhance copying degree, demonstrating that high-copying responses achieve superior contextual faithfulness and hallucination control. These approaches enable a fully automated pipeline that transforms generated responses into high-copying preference data for training CopyPasteLLM. On FaithEval, ConFiQA and PubMedQA, CopyPasteLLM achieves best performance in both counterfactual and original contexts, remarkably with 12.2% to 24.5% accuracy improvements on FaithEval over the best baseline, while requiring only 365 training samples -- 1/50th of baseline data. To elucidate CopyPasteLLM's effectiveness, we propose the Context-Parameter Copying Capturing algorithm. Interestingly, this reveals that CopyPasteLLM recalibrates reliance on internal parametric knowledge rather than external knowledge during generation. All codes are available at https://github.com/longyongchao/CopyPasteLLM

Microscopy techniques generate vast amounts of complex image data that in principle can be used to discover simpler, interpretable, and parsimonious forms to reveal the underlying physical structures, such as elementary building blocks in molecular systems or order parameters and phases in crystalline materials. Variational Autoencoders (VAEs) provide a powerful means of constructing such low-dimensional representations, but their performance heavily depends on multiple non-myopic design choices, which are often optimized through trial-and-error and empirical analysis. To enable automated and unbiased optimization of VAE workflows, we investigated reward-based strategies for evaluating latent space representations. Using Piezoresponse Force Microscopy data as a model system, we examined multiple policies and reward functions that can serve as a foundation for automated optimization. Our analysis shows that approximating the latent space with Gaussian Mixture Models (GMM) and Bayesian Gaussian Mixture Models (BGMM) provides a strong basis for constructing reward functions capable of estimating model efficiency and guiding the search for optimal parsimonious representations.

Foundational models based on the transformer architecture are currently the state-of-the-art in general language modeling, as well as in scientific areas such as material science and climate. However, training and deploying these models is computationally challenging as the time and space complexity has a quadratic relation to the input sequence length. Several efforts exploring efficient computational paradigms and model architectures to address these limitations have been made. In this work, we explore spiking neural networks (SNNs) to design transformer models. A challenge in training large-scale SNNs, using existing surrogate learning methods is inefficient and time-consuming. On the other hand, techniques to convert existing transformer-based models to their SNN equivalent are not scalable, as achieving optimal performance comes at the cost of a large number of spike time-steps, i.e. increased latency. To address this, we propose NeurTransformer, a methodology for designing transformer-based SNN for inference using a supervised fine-tuning approach with existing conversion methods. The proposed methodology works by: (1) replacing the self-attention mechanism with a spike-based self-attention (SSA), (2) converting the feed-forward block of the trained transformer model to its equivalent SNN, and (3) fine-tuning the SSA block using SNN-based surrogate learning algorithms. We benchmark the proposed methodology and demonstrate its accuracy and scalability using three variants of the GPT-2 model of increasing model size. We observe that the converted GPT-2 small models demonstrate a 5-12% loss in cosine similarity and a 9.7% reduction in perplexity. Finally, we demonstrate the energy efficiency of the SSA block compared to the ASA block and show between 64.71% and 85.28% reductions in estimated energy consumption when implementing the self-attention mechanism on a digital hardware.

Strain gradients, ubiquitous in flexible devices and epitaxial nanostructures, are a major blind spot for thermal transport in \b{eta}-Ga2O3. We establish that strain gradient unlocks a thermal conductivity (k) suppression mechanism fundamentally more potent than uniform strain: moderate uniaxial gradients (0.6%/nm) suppress k by 32-37% (27-30%) in thin films (nanowires), intensifying to 43.3% with biaxial gradients. This reduction far exceeds that from equivalent uniform strain and surpasses benchmark materials like silicon and BAs. Critically, a surprising decoupling emerges: while 3% uniform strain alters thermal anisotropy by ~25%, strain gradient strongly suppresses k with preserving this ratio. Mechanistically, strain gradients-induced symmetry breaking and enhanced mode coupling anisotropically activate forbidden scattering channels, making gradient-driven scattering dominant over intrinsic phonon scattering below 6.25 THz. These findings redefine non-uniform strain from a parasitic flaw into a powerful design tool for engineering thermal isolation and heat flux in next-generation flexible and high-power \b{eta}-Ga2O3 electronics.

Large Language Models (LLMs) have garnered significant attention for several years now. Recently, their use as independently reasoning agents has been proposed. In this work, we test the potential of such agents for knowledge discovery in materials science. We repurpose LangGraph's tool functionality to supply agents with a black box function to interrogate. In contrast to process optimization or performing specific, user-defined tasks, knowledge discovery consists of freely exploring the system, posing and verifying statements about the behavior of this black box, with the sole objective of generating and verifying generalizable statements. We provide proof of concept for this approach through a children's parlor game, demonstrating the role of trial-and-error and persistence in knowledge discovery, and the strong path-dependence of results. We then apply the same strategy to show that LLM agents can explore, discover, and exploit diverse chemical interactions in an advanced Atomic Layer Processing reactor simulation using intentionally limited probe capabilities without explicit instructions.

Considerable effort continues to be devoted to the exploration of next-generation high-\k{appa} materials that combine a high dielectric constant with a wide band gap. However, machine learning (ML)-based virtual screening has remained challenging, primarily due to the low accuracy in predicting the ionic contribution to the dielectric tensor, which dominates the dielectric performance of high-\k{appa} materials. We here propose a joint ML model that predicts Born effective charges using an equivariant graph neural network, and phonon properties using a highly accurate pretrained ML potential. The ionic dielectric tensor is then computed analytically from these quantities. This approach significantly improves the accuracy of ionic contribution. Using the proposed model, we successfully identified 38 novel high-\k{appa} oxides from a screening pool of over 8,000 candidates.

Accurate prediction of surface energies and stabilities is essential for materials design, yet first-principles calculations remain computationally expensive and most existing interatomic potentials are trained only on bulk systems. Here, we demonstrate that fine-tuning foundation machine learning potentials (MLPs) significantly improves both computational efficiency and predictive accuracy for surface modeling. While existing universal interatomic potentials (UIPs) have been solely trained and validated on bulk datasets, we extend their applicability to complex and scientifically significant unary, binary, and ternary surface systems. We systematically compare models trained from scratch, zero-shot inference, conventional fine-tuning, and multi-head fine-tuning approach that enhances transferability and mitigates catastrophic forgetting. Fine-tuning consistently reduces prediction errors with orders-of-magnitude fewer training configurations, and multi-head fine-tuning delivers robust and generalizable predictions even for materials beyond the initial training domain. These findings offer practical guidance for leveraging pre-trained MLPs to accelerate surface modeling and highlight a scalable path toward data-efficient, next-generation atomic-scale simulations in computational materials science.

Oxide Li-conducting solid-state electrolytes (SSEs) offer excellent chemical and thermal stability but typically exhibit lower ionic conductivity than sulfides and chlorides. This motivates the search for new oxide materials with enhanced conductivity. Crystal structure prediction is a powerful approach for identifying such candidates. However, the structural complexity of oxide SSEs, often involving unit cells with more than 100 atoms, presents significant challenges for conventional methods. In this study, we introduce TOPIC, a structure prediction algorithm that reduces configurational complexity by enforcing corner-sharing (CS) bond topology constraints. We demonstrate that TOPIC successfully reproduces the ground-state and metastable structures of known oxide SSEs, including LiTa$_2$PO$_8$ and Li$_7$La$_3$Zr$_2$O$_{12}$, which contain up to about 200 atoms per unit cell. By combining this approach with a pretrained machine-learning interatomic potential, we systematically screen quaternary oxide compositions and identify 92 promising candidates with CS frameworks. In particular, Li$_4$Hf$_2$Si$_3$O$_{12}$, which corresponds to the ground state at its composition, exhibits an ionic conductivity of 14 mS cm$^{-1}$, a hull energy of 21 meV atom$^{-1}$, and a band gap of 6.5 eV. Through our investigation, we identify the Li ratio as one of the key factors determining the stability of CS structures. Overall, our approach provides a practical and scalable pathway for discovering high-performance oxide solid electrolytes in previously unexplored chemical spaces.

Generative AI poses both opportunities and risks for solving inverse design problems in the sciences. Generative tools provide the ability to expand and refine a search space autonomously, but do so at the cost of exploring low-quality regions until sufficiently fine tuned. Here, we propose a queue prioritization algorithm that combines generative modeling and active learning in the context of a distributed workflow for exploring complex design spaces. We find that incorporating an active learning model to prioritize top design candidates can prevent a generative AI workflow from expending resources on nonsensical candidates and halt potential generative model decay. For an existing generative AI workflow for discovering novel molecular structure candidates for carbon capture, our active learning approach significantly increases the number of high-quality candidates identified by the generative model. We find that, out of 1000 novel candidates, our workflow without active learning can generate an average of 281 high-performing candidates, while our proposed prioritization with active learning can generate an average 604 high-performing candidates.

Recent advances in structure-based protein design have accelerated de novo binder generation, yet interfaces on large domains or spanning multiple domains remain challenging due to high computational cost and declining success with increasing target size. We hypothesized that protein folding neural networks (PFNNs) operate in a ``local-first'' manner, prioritizing local interactions while displaying limited sensitivity to global foldability. Guided by this hypothesis, we propose an epitope-only strategy that retains only the discontinuous surface residues surrounding the binding site. Compared to intact-domain workflows, this approach improves in silico success rates by up to 80% and reduces the average time per successful design by up to forty-fold, enabling binder design against previously intractable targets such as ClpP and ALS3. Building on this foundation, we further developed a tailored pipeline that incorporates a Monte Carlo-based evolution step to overcome local minima and a position-specific biased inverse folding step to refine sequence patterns. Together, these advances not only establish a generalizable framework for efficient binder design against structurally large and otherwise inaccessible targets, but also support the broader ``local-first'' hypothesis as a guiding principle for PFNN-based design.

The inverse design of materials with specific desired properties, such as high-temperature superconductivity, represents a formidable challenge in materials science due to the vastness of chemical and structural space. We present a guided diffusion framework to accelerate the discovery of novel superconductors. A DiffCSP foundation model is pretrained on the Alexandria Database and fine-tuned on 7,183 superconductors with first principles derived labels. Employing classifier-free guidance, we sample 200,000 structures, which lead to 34,027 unique candidates. A multistage screening process that combines machine learning and density functional theory (DFT) calculations to assess stability and electronic properties, identifies 773 candidates with DFT-calculated $T_\mathrm{c}>5$ K. Notably, our generative model demonstrates effective property-driven design. Our computational findings were validated against experimental synthesis and characterization performed as part of this work, which highlighted challenges in sparsely charted chemistries. This end-to-end workflow accelerates superconductor discovery while underscoring the challenge of predicting and synthesizing experimentally realizable materials.

Metal inorganic-organic complex (MIOC) crystals are a new category of hybrid glass formers. However, the glass-forming compositions of MIOC crystals are limited due to lack of both a general design principle for such compositions and a deep understanding of the structure and formation mechanism for MIOC glasses. This work reports a general approach for synthesizing glass-forming MIOC crystals. In detail, the principle of this approach is based on the creation of hydrogen-bonded structural network by substituting acid anions for imidazole or benzimidazole ligands in the tetrahedral units of zeolitic imidazolate framework crystals. By tuning the metal centers, anions, and organic ligands of MIOCs, supramolecular unit structures can be designed to construct supramolecular networks and thereby enable property modulation. Furthermore, mixed-ligand synthesis yielded a mixed-crystal system in which the glass-transition temperature (Tg) can be linearly tuned from 282 K to 360 K through gradual substitution of benzimidazole for imidazole. Interestingly, upon vitrification, MIOCs were observed to undergo reorganization of hydrogen-bonded networks, with retention of tetrahedral units, short-range disorder, and the freezing of multiple conformations. This work offers a new strategy to systematically expand the glass-forming compositional range of MIOCs and to develop functional MIOC glasses.

We present PoseDiff, a conditional diffusion model that unifies robot state estimation and control within a single framework. At its core, PoseDiff maps raw visual observations into structured robot states-such as 3D keypoints or joint angles-from a single RGB image, eliminating the need for multi-stage pipelines or auxiliary modalities. Building upon this foundation, PoseDiff extends naturally to video-to-action inverse dynamics: by conditioning on sparse video keyframes generated by world models, it produces smooth and continuous long-horizon action sequences through an overlap-averaging strategy. This unified design enables scalable and efficient integration of perception and control. On the DREAM dataset, PoseDiff achieves state-of-the-art accuracy and real-time performance for pose estimation. On Libero-Object manipulation tasks, it substantially improves success rates over existing inverse dynamics modules, even under strict offline settings. Together, these results show that PoseDiff provides a scalable, accurate, and efficient bridge between perception, planning, and control in embodied AI. The video visualization results can be found on the project page: https://haozhuo-zhang.github.io/PoseDiff-project-page/.

Solving inverse problems with diffusion models has shown promise in tasks such as image restoration. A common approach is to formulate the problem in a Bayesian framework and sample from the posterior by combining the prior score with the likelihood score. Since the likelihood term is often intractable, estimators like DPS, DMPS, and $π$GDM are widely adopted. However, these methods rely on a fixed, manually tuned scale to balance prior and likelihood contributions. Such a static design is suboptimal, as the ideal balance varies across timesteps and tasks, limiting performance and generalization. To address this issue, we propose SAIP, a plug-and-play module that adaptively refines the scale at each timestep without retraining or altering the diffusion backbone. SAIP integrates seamlessly into existing samplers and consistently improves reconstruction quality across diverse image restoration tasks, including challenging scenarios.

Solid-melt interfaces play a pivotal role in governing crystal growth and metal-mediated epitaxy of gallium nitride (GaN) and other semiconductor materials. Using atomistic simulations based on machine-learning interatomic potentials (MLIPs), we uncover that multiple layers of Ga atoms at the GaN-Ga melt interface form structurally ordered and electronically charged configurations that are critical for the growth kinetics of GaN. These ordered layers modulate the free energy landscape (FEL) for N adsorption and substantially reduce the migration barriers for N at the interface compared to a clean GaN surface. Leveraging these interfacial energetics, kinetic Monte Carlo (KMC) simulations reveal that GaN growth follows a diffusion-controlled, layer-by-layer mechanism, with the FEL for N adsorption emerging as the rate-limiting factor. By incorporating facet-specific FELs and the diffusivity/solubility of N in Ga melt, we develop a predictive, fitting-free transport model that estimates facet-dependent growth rates in the range of ~0.01 to 0.04 nm/s, in agreement with experimental growth rates observed in GaN nanoparticles synthesized by Ga-mediated molecular beam epitaxy (MBE). This multiscale framework offers a generalizable and quantitative approach to link atomic-scale ordering and interfacial energetics to macroscopic phenomena, providing actionable insights for the rational design of metal-mediated epitaxial processes.

Gravity inversion is the problem of estimating subsurface density distributions from observed gravitational field data. We consider the two-dimensional (2D) case, in which recovering density models from one-dimensional (1D) measurements leads to an underdetermined system with substantially more model parameters than measurements, making the inversion ill-posed and non-unique. Recent advances in machine learning have motivated data-driven approaches for gravity inversion. We first design a convolutional neural network (CNN) trained to directly map gravity anomalies to density fields, where a customized data structure is introduced to enhance the inversion performance. To further investigate generative modeling, we employ Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs), reformulating inversion as a latent-space optimization constrained by the forward operator. In addition, we assess whether classical iterative solvers such as Gradient Descent (GD), GMRES, LGMRES, and a recently proposed Improved Conjugate Gradient (ICG) method can refine CNN-based initial guesses and improve inversion accuracy. Our results demonstrate that CNN inversion not only provides the most reliable reconstructions but also significantly outperforms previously reported methods. Generative models remain promising but unstable, and iterative solvers offer only marginal improvements, underscoring the persistent ill-posedness of gravity inversion.