arXiv subjects

Rafael Gomez-Bombarelli

Publications and source records attributed to Rafael Gomez-Bombarelli.

At least 19 recordsLinked to original sources

arXiv2026

AtomSteer: Test-Time Constraint Steering for Atomistic Generative Models

Diffusion and flow matching models offer scalable approaches to generating crystalline materials, molecular conformers, and protein-ligand complexes. However, satisfying crystallographic symmetry, structural geometry, and accurate stereochemistry remains challenging. Satisfying these requirements calls for corrections that preserve the geometry already captured by the pretrained model. In this work, we propose AtomSteer, a test-time inference framework for steering pretrained atomistic generative models with structural constraints. During sampling, AtomSteer corrects intermediate structure predictions with bounded local projections, graph-rigid stereochemical transformations, and projections over periodic symmetries, accepting updates that improve constraint satisfaction while limiting changes to the rest of the structure. We demonstrate AtomSteer across three pretrained backbones, FlowMM, ET-Flow, and Boltz-2, spanning de novo crystal generation and crystal structure prediction, molecular conformer generation, and protein-ligand complex structure prediction. AtomSteer improves space-group distribution coverage in de novo crystal generation and improves target space-group recovery in crystal structure prediction. AtomSteer improves molecular validity and E/Z stereochemical accuracy while retaining conformer ensemble fidelity. For protein-ligand complexes, AtomSteer reduces ligand E/Z errors relative to experimental reference structures while achieving competitive docking success and PoseBusters validity. Our method provides a unified framework for adding chemical and crystallographic control to existing atomistic generative models without retraining or architectural changes.

cs.LG↗

arXiv2026

PXRD-Bench: How Close Are We to Determining Crystal Structures from Powder Diffraction?

Determining crystal structures from powder X-ray diffraction (PXRD) would enable routine characterization of newly synthesized materials, including those absent from reference databases. Generative models offer a promising approach, but existing evaluations remain fragmented and largely assume single-phase samples with known composition. To fill this gap, we introduce PXRD-Bench, a unified benchmark that extends evaluation to unknown per-element atom counts and recovery of both structures in two-phase mixtures. It provides approximately 4.1 million training structures with simulated PXRD, tests generalization to unseen compositions and structural prototypes, and evaluates recovery from experimental PXRD after excluding structurally matching training entries. To provide a strong reference baseline, we also introduce Bragglite, a PXRD-to-structure model developed through a systematic study of PXRD modeling choices. It achieves a 44.2\% Top-1 match rate on experimental RRUFF patterns with known composition, 24.6\% above the strongest prior baseline. Further analysis yields several key findings: some models fail to benefit from experimental PXRD, while generalization to unseen structural prototypes and recovery from two-phase mixtures remain challenging. Together, our benchmark, systematic studies, and Bragglite clarify the capabilities and limitations of current PXRD-to-structure models.

cs.AI↗

arXiv2026

Drude, Where's My Charge? End-to-End Learning of Polarizable Drude Force Fields for Polymer Electrolytes

Molecular dynamics simulations of solid polymer electrolytes require force fields that are fast enough to converge ion transport over hundreds of nanoseconds, yet accurately describe the ion-polymer interactions that govern it. Fixed-charge force fields meet the first requirement but neglect electronic polarization, an effect often approximated by scaling ionic charges using empirically chosen factors. Here we present AutoDrude, a differentiable framework that parametrizes polarizable Drude force fields directly from DFT energies, forces, and dipoles, with no empirical tuning and at only 1.20 times the cost of a fixed-charge simulation. For PEO-LiTFSI, the dipole-constrained (hard-constraint) AutoDrude model reproduces the experimental ionic conductivity as accurately as the best charge-scaled model (17% versus 16% mean absolute percentage error). We compare the polarizable forcefield to fixed charge ones models by fitting a fixed charge model. Unlike the charge-scaled models tested here, however, AutoDrude also reproduces the dilute-limit cation transference number, its increase with salt concentration, and the absence of Li-TFSI contact ion pairs at low concentration. A penalty-based (soft-constraint) variant yields a closer fit to the DFT reference yet underestimates the conductivity by up to an order of magnitude, demonstrating that closer agreement with DFT training data alone does not guarantee accurate ion transport. These results indicate that although charge scaling can reproduce ionic conductivity the explicit inclusion of polarization is necessary to capture the concentration dependence of cation transport in PEO-LiTFSI.

cond-mat.mtrl-sci↗

arXiv2026

SNAP-FM: Sparse Nonlinear Accelerated Projection for Physics-Constrained Generative Modeling

Generative models have emerged as scalable surrogates for physical simulation, yet they offer no guarantee that their outputs respect the conservation laws, boundary conditions, and nonlinear invariants that govern the underlying physics. Constrained sampling closes this gap, enforcing such constraints exactly at inference time without retraining, but at a computational cost: projection, correction and trajectory-optimization steps are repeated during sampling, with these steps becoming expensive for nonlinear constraints. Standard ML frameworks exacerbate this: their dense tensor algebra and limited sparse solver composability obscure the structure that physical constraints naturally induce, making efficient batched nonlinear optimization difficult to realize in practice. We address this bottleneck by exploiting the structure that sample-wise batching and local PDE couplings induce in the projection subproblems -- namely, block-sparse Jacobian and KKT systems -- exposing this structure using ExaModels.jl and solving the resulting sparse nonlinear programs with MadNLP.jl and GPU sparse factorization. Applied to Physics-Constrained Flow Matching (PCFM), on PDE benchmarks with linear, nonlinear, one-dimensional, and two-dimensional constraints, this approach accelerates nonlinear constraint projection while maintaining constraint satisfaction. These results show that sparse GPU nonlinear optimization is a practical foundation for constrained generative sampling in scientific machine learning.

cs.LG↗

arXiv2026

Reinforcement Learning with Verifiable Physics: Post-training LLMs with Continuous Rewards

Partial differential equations (PDEs) are foundational to modeling in science and engineering, but constructing reliable numerical solvers remains labor-intensive, demanding expert knowledge of discretization schemes, stability conditions, and boundary treatments. Recent work has begun to frame PDE solving as a code-generation task for large language models (LLMs), yet existing approaches operate primarily at inference time: relying on prompting, debugging, self-refinement, and test-time scaling rather than adapting the model itself. In parallel, reinforcement learning with verifiable rewards has emerged as a post-training paradigm for code and math reasoning, but its verifiers are typically binary: a compiler runs, or a test passes. Such signals discard the graded structure of scientific correctness, where two solvers may both execute and yet differ in solution accuracy by orders of magnitude. In this work, we introduce RLVP: Reinforcement Learning with Verifiable Physics, an RL post-training framework for multi-PDE solver code generation. RLVP addresses this verifiability gap with a hybrid verifier: hard program-validity checks ensure executability, while continuous physics rewards score function-space accuracy and PDE-residual consistency. A single policy is post-trained across diverse PDE families spanning hyperbolic, parabolic, elliptic, and incompressible-flow systems. RLVP improves over both pre-trained and supervised-only baselines on PDE benchmarks, and shows zero-shot improvement transfer to held-out PDEs. We show that a smaller LLM post-trained with RLVP can outperform prompting a frontier model on in-distribution PDE solver generation. The trained policy shows evidence of compositionality in numerical motifs: it recombines stencils, time-stepping schemes, and boundary-handling primitives learned from the PDEs used in training into generated solvers for unseen PDE problems.

cs.LG↗

arXiv2026

Zatom-1: Towards a Multimodal Foundation Model for 3D Molecules and Materials

General-purpose 3D modeling in chemistry encompasses molecules and materials, requiring both generative and predictive capabilities. However, most existing AI approaches are optimized for a single domain (molecules or materials) and a single task (generation or prediction), which limits representation sharing and transfer. We introduce Zatom-1, a cross-domain, general-purpose model architecture that unifies generative and predictive learning of 3D molecules and materials. Zatom-1 is a deliberately simplified Transformer trained with a multimodal flow matching objective that jointly models discrete atom types and continuous 3D geometries. This approach supports scalable pretraining with predictable gains as model capacity increases, while enabling fast and stable sampling. We use cross-domain generative pretraining as a universal initialization for downstream multi-task prediction of properties, energies, and forces. Empirically, Zatom-1 outperforms or competes with specialized baselines on both multi-task generative and predictive benchmarks in data-controlled settings, while improving generative inference speed by more than an order of magnitude. Our experiments demonstrate positive predictive transfer between data domains from joint generative pretraining: modeling materials during generative pretraining improves molecular property prediction accuracy. Open-source code and model weights are freely available at https://github.com/Zatom-AI/zatom.

cs.LG↗

arXiv2026

Self-Conditioned Denoising for Atomistic Representation Learning

The success of large-scale pretraining in NLP and computer vision has catalyzed growing efforts to develop analogous foundation models for the physical sciences. However, pretraining strategies using atomistic data remain underexplored. To date, large-scale supervised pretraining on DFT force-energy labels has provided the strongest performance gains to downstream property prediction, out-performing existing methods of self-supervised learning (SSL) which remain limited to ground-state geometries, and/or single domains of atomistic data. We address these shortcomings with Self-Conditioned Denoising (SCD), a backbone-agnostic reconstruction objective that utilizes self-embeddings for conditional denoising across any domain of atomistic data, including small molecules, proteins, periodic materials, and 'non-equilibrium' geometries. When controlled for backbone architecture and pretraining dataset, SCD significantly outperforms previous SSL methods on downstream benchmarks and matches or exceeds the performance of supervised force-energy pretraining. We show that a small, fast GNN pretrained by SCD can achieve competitive or superior performance to larger models pretrained on significantly larger labeled or unlabeled datasets, across tasks in multiple domains. Our code is available at: https://github.com/TyJPerez/SelfConditionedDenoisingAtoms

cs.LG↗

arXiv2026

The Future of Artificial Intelligence and the Mathematical and Physical Sciences (AI+MPS)

This community paper developed out of the NSF Workshop on the Future of Artificial Intelligence (AI) and the Mathematical and Physics Sciences (MPS), which was held in March 2025 with the goal of understanding how the MPS domains (Astronomy, Chemistry, Materials Research, Mathematical Sciences, and Physics) can best capitalize on, and contribute to, the future of AI. We present here a summary and snapshot of the MPS community's perspective, as of Spring/Summer 2025, in a rapidly developing field. The link between AI and MPS is becoming increasingly inextricable; now is a crucial moment to strengthen the link between AI and Science by pursuing a strategy that proactively and thoughtfully leverages the potential of AI for scientific discovery and optimizes opportunities to impact the development of AI by applying concepts from fundamental science. To achieve this, we propose activities and strategic priorities that: (1) enable AI+MPS research in both directions; (2) build up an interdisciplinary community of AI+MPS researchers; and (3) foster education and workforce development in AI for MPS researchers and students. We conclude with a summary of suggested priorities for funding agencies, educational institutions, and individual researchers to help position the MPS community to be a leader in, and take full advantage of, the transformative potential of AI+MPS.

cs.AI↗

arXiv2026

PackFlow: Generative Molecular Crystal Structure Prediction via Reinforcement Learning Alignment

Organic molecular crystals underpin technologies ranging from pharmaceuticals to organic electronics, yet predicting solid-state packing of molecules remains challenging because candidate generation is combinatorial and stability is only resolved after costly energy evaluations. Here we introduce PackFlow, a flow matching framework for molecular crystal structure prediction (CSP) that generates heavy-atom crystal proposals by jointly sampling Cartesian coordinates and unit-cell lattice parameters given a molecular graph. This lattice-aware generation interfaces directly with downstream relaxation and lattice-energy ranking, positioning PackFlow as a scalable proposal engine within standard CSP pipelines. To explicitly steer generation toward physically favourable regions, we propose physics alignment, a reinforcement learning post-training stage that uses machine-learned interatomic potential energies and forces as stability proxies. Physics alignment improves physical validity without altering inference-time sampling. We validate PackFlow's performance against heuristic baselines through two distinct evaluations. First, on a broad unseen set of molecular systems, we demonstrate superior candidate generation capability, with proposals exhibiting greater structural similarity to experimental polymorphs. Second, we assess the full end-to-end workflow on two unseen CSP blind-test case studies, including relaxation and lattice-energy analysis. In both settings, PackFlow outperforms heuristics-based methods by concentrating probability mass in low-energy basins, yielding candidates that relax into lower-energy minima and offering a practical route to amortize the relax-and-rank bottleneck.

physics.chem-ph↗

arXiv2025

Thermal Equilibrium Vacancy Concentration in an Alloy with Chemical Short-Range Order

The equilibrium vacancy concentration in multi-principal element alloys remains a controversial and nontrivial subject, primarily because of chemical complexity and chemical short-range order (CSRO). Here we derive an exact expression that is amenable to atomistic calculations, using multiple perspectives. We applied this expression to equiatomic CrCoNi alloys in the face-centered cubic structure. The derived equilibrium vacancy concentration is used in our recent work, which predicts the chemical short-range order formation timescale consistent with experimental observation. The results demonstrate the practical utility of the approach for predicting equilibrium vacancy concentrations in compositionally complex alloys.

cond-mat.mtrl-sci↗

arXiv2025

Deep Learning of Mean First Passage Time Scape: Chemical Short-Range Order and Kinetics of Diffusive Relaxation

Processes slow compared to atomic vibrations pose significant challenges in atomistic simulations, particularly for phenomena such as diffusive relaxations and phase transitions, where repeated crossings and the shear number of thermally activated transitions make direct numerical simulations impossible. We present a computational framework that captures atomic-scale diffusive relaxation over extended timescales by learning the mean first passage time (MFPT) with a deep neural network. The model is trained via a self-consistent recursive formulation based on the Markovian assumption, relying solely on local residence times and transition probabilities between neighboring states. Furthermore, we leverage deep reinforcement learning (DRL)-accelerated atomistic simulations to expedite the identification of thermodynamic equilibrium and the generation of accurate physical transition probabilities. Applied to vacancy-mediated chemical short-range order (SRO) evolution in equiatomic CrCoNi, our method uncovers disorder-to-order transition timescales in quantitative agreement with experimental measurements. By bridging the gap between simulation and experiment, our approach extends atomistic modeling to previously inaccessible timescales and offers a predictive tool for navigating process-structure-property relationships.

cond-mat.mtrl-sci↗

arXiv2025

The Loss Landscape of Powder X-Ray Diffraction-Based Structure Optimization Is Too Rough for Gradient Descent

Solving crystal structures from powder X-ray diffraction (XRD) is a central challenge in materials characterization. In this work, we study the powder XRD-to-structure mapping using gradient descent optimization, with the goal of recovering the correct structure from moderately distorted initial states based solely on XRD similarity. We show that commonly used XRD similarity metrics result in a highly non-convex landscape, complicating direct optimization. Constraining the optimization to the ground-truth crystal family significantly improves recovery, yielding higher match rates and increased mutual information and correlation scores between structural similarity and XRD similarity. Nevertheless, the landscape may remain non-convex along certain symmetry axes. These findings suggest that symmetry-aware inductive biases could play a meaningful role in helping learning models navigate the inverse mapping from diffraction to structure.

cond-mat.mtrl-sci↗

arXiv2025

Physics-Constrained Flow Matching: Sampling Generative Models with Hard Constraints

Deep generative models have recently been applied to physical systems governed by partial differential equations (PDEs), offering scalable simulation and uncertainty-aware inference. However, enforcing physical constraints, such as conservation laws (linear and nonlinear) and physical consistencies, remains challenging. Existing methods often rely on soft penalties or architectural biases that fail to guarantee hard constraints. In this work, we propose Physics-Constrained Flow Matching (PCFM), a zero-shot inference framework that enforces arbitrary nonlinear constraints in pretrained flow-based generative models. PCFM continuously guides the sampling process through physics-based corrections applied to intermediate solution states, while remaining aligned with the learned flow and satisfying physical constraints. Empirically, PCFM outperforms both unconstrained and constrained baselines on a range of PDEs, including those with shocks, discontinuities, and sharp features, while ensuring exact constraint satisfaction at the final solution. Our method provides a flexible framework for enforcing hard constraints in both scientific and general-purpose generative models, especially in applications where constraint satisfaction is essential.

cs.LG↗

arXiv2025

DiffSyn: A Generative Diffusion Approach to Materials Synthesis Planning

The synthesis of crystalline materials, such as zeolites, remains a significant challenge due to a high-dimensional synthesis space, intricate structure-synthesis relationships and time-consuming experiments. Considering the one-to-many relationship between structure and synthesis, we propose DiffSyn, a generative diffusion model trained on over 23,000 synthesis recipes spanning 50 years of literature. DiffSyn generates probable synthesis routes conditioned on a desired zeolite structure and an organic template. DiffSyn achieves state-of-the-art performance by capturing the multi-modal nature of structure-synthesis relationships. We apply DiffSyn to differentiate among competing phases and generate optimal synthesis routes. As a proof of concept, we synthesize a UFI material using DiffSyn-generated synthesis routes. These routes, rationalized by density functional theory binding energies, resulted in the successful synthesis of a UFI material with a high Si/Al$_{\text{ICP}}$ of 19.0, which is expected to improve thermal stability and is higher than that of any previously recorded.

cond-mat.mtrl-sci↗

arXiv2025

Known Unknowns: Out-of-Distribution Property Prediction in Materials and Molecules

Discovery of high-performance materials and molecules requires identifying extremes with property values that fall outside the known distribution. Therefore, the ability to extrapolate to out-of-distribution (OOD) property values is critical for both solid-state materials and molecular design. Our objective is to train predictor models that extrapolate zero-shot to higher ranges than in the training data, given the chemical compositions of solids or molecular graphs and their property values. We propose using a transductive approach to OOD property prediction, achieving improvements in prediction accuracy. In particular, the True Positive Rate (TPR) of OOD classification of materials and molecules improved by 3x and 2.5x, respectively, and precision improved by 2x and 1.5x compared to non-transductive baselines. Our method leverages analogical input-target relations in the training and test sets, enabling generalization beyond the training target support, and can be applied to any other material and molecular tasks.

cs.LG↗

arXiv2024

Univariate Conditional Variational Autoencoder for Morphogenic Patterns Design in Frontal Polymerization-Based Manufacturing

Under some initial and boundary conditions, the rapid reaction-thermal diffusion process taking place during frontal polymerization (FP) destabilizes the planar mode of front propagation, leading to spatially varying, complex hierarchical patterns in thermoset polymeric materials. Although modern reaction-diffusion models can predict the patterns resulting from unstable FP, the inverse design of patterns, which aims to retrieve process conditions that produce a desired pattern, remains an open challenge due to the non-unique and non-intuitive mapping between process conditions and manufactured patterns. In this work, we propose a probabilistic generative model named univariate conditional variational autoencoder (UcVAE) for the inverse design of hierarchical patterns in FP-based manufacturing. Unlike the cVAE, which encodes both the design space and the design target, the UcVAE encodes only the design space. In the encoder of the UcVAE, the number of training parameters is significantly reduced compared to the cVAE, resulting in a shorter training time while maintaining comparable performance. Given desired pattern images, the trained UcVAE can generate multiple process condition solutions that produce high-fidelity hierarchical patterns.

physics.comp-ph↗

arXiv2024

Enhanced sampling of robust molecular datasets with uncertainty-based collective variables

Generating a data set that is representative of the accessible configuration space of a molecular system is crucial for the robustness of machine learned interatomic potentials (MLIP). However, the complexity of molecular systems, characterized by intricate potential energy surfaces (PESs) with numerous local minima and energy barriers, presents a significant challenge. Traditional methods of data generation, such as random sampling or exhaustive exploration, are either intractable or may not capture rare, but highly informative configurations. In this study, we propose a method that leverages uncertainty as the collective variable (CV) to guide the acquisition of chemically-relevant data points, focusing on regions of the configuration space where ML model predictions are most uncertain. This approach employs a Gaussian Mixture Model-based uncertainty metric from a single model as the CV for biased molecular dynamics simulations. The effectiveness of our approach in overcoming energy barriers and exploring unseen energy minima, thereby enhancing the data set in an active learning framework, is demonstrated on the alanine dipeptide benchmark system.

cs.LG↗

arXiv2023

Forces are not Enough: Benchmark and Critical Evaluation for Machine Learning Force Fields with Molecular Simulations

Molecular dynamics (MD) simulation techniques are widely used for various natural science applications. Increasingly, machine learning (ML) force field (FF) models begin to replace ab-initio simulations by predicting forces directly from atomic structures. Despite significant progress in this area, such techniques are primarily benchmarked by their force/energy prediction errors, even though the practical use case would be to produce realistic MD trajectories. We aim to fill this gap by introducing a novel benchmark suite for learned MD simulation. We curate representative MD systems, including water, organic molecules, a peptide, and materials, and design evaluation metrics corresponding to the scientific objectives of respective systems. We benchmark a collection of state-of-the-art (SOTA) ML FF models and illustrate, in particular, how the commonly benchmarked force accuracy is not well aligned with relevant simulation metrics. We demonstrate when and how selected SOTA methods fail, along with offering directions for further improvement. Specifically, we identify stability as a key metric for ML models to improve. Our benchmark suite comes with a comprehensive open-source codebase for training and simulation with ML FFs to facilitate future work.

physics.comp-ph↗