arXiv subjects

Maya M. Martirossyan

Publications and source records attributed to Maya M. Martirossyan.

8 recordsLinked to original sources

arXiv2026

Riemannian Flow Models with Reinforcement Learning for Molecular Crystal Structure Prediction

Crystal structure governs material properties, making crystal structure prediction (CSP) a fundamental problem in materials science. Generative models are a promising approach for solving this problem, but the prevalence of polymorphism, coupled with large unit cells and complex packing geometry, makes the molecular CSP task challenging for existing models. To address this, we introduce Coarse-Grained Open Materials Generation (CG-OMatG), an equivariant Riemannian flow-based generative model. CG-OMatG predicts molecular crystal structures \textit{via} a coarse-grained, hierarchical representation. CG-OMatG treats molecules as rigid bodies---performing both inter- and intra-molecular message passing to construct a geometric representation for molecular packings---and learns to reconstruct molecule centroid positions, orientations, and lattice parameters, conditioned on chemical species and conformer geometry. We train the model on subsets of the Open Molecular Crystals (OMC25) and Cambridge Structural Database (CSD) datasets. Further, we fine-tune the model \textit{via} policy gradient reinforcement learning to steer the model towards generating low-energy candidate structures. We validate the generated structures on the CSP blind test benchmark, assessing agreement with experimentally determined crystals using COMPACK packing-similarity analysis. CG-OMatG exhibits strong performance for generative molecular crystal structure prediction, paving the way for accelerated polymorph screening and organic solid-state materials discovery.

cs.LG↗

arXiv2026

MolCryst-MLIPs: A Machine-Learned Interatomic Potentials Database for Molecular Crystals

We present an open Molecular Crystal (MC) database of Machine-Learned Interatomic Potentials (MLIP) called MolCryst-MLIPs. The first release comprises fine-tuned MACE models for nine molecular crystal systems---Benzamide, Benzoic acid, Coumarin, Durene, Isonicotinamide, Nicotinic acid , Niacinamide, Pyrazinamide, and Resorcinol---developed using the Automated Machine Learning Pipeline (AMLP), which streamlines the entire MLIP development workflow, from reference data generation to model training and validation, into a reproducible and user-friendly pipeline. Models are fine-tuned from the MACE-MH-1 foundation model omol head), yielding a mean energy MAE of 0.141 kJ/mol/atom and a mean force MAE of 0.648 kJ/mol/Angstrom across all systems. Benchmarked against three state-of-the-art foundation models on the DFT-labelled polymorph set, only the fine-tuned models resolve the polymorphic energy landscape. Dynamical stability and structural integrity, as assessed through energy conservation, P2 orientational order parameters, and radial distribution functions, are evaluated using molecular dynamics simulations. The released models and datasets constitute a growing open database of validated MLIPs, ready for production MD simulations of molecular crystal polymorphism across the polymorphic landscape of each target compound under different thermodynamic conditions.

cs.LG↗

arXiv2026

MolCrystalFlow: Molecular Crystal Structure Prediction via Flow Matching

Molecular crystal structure prediction represents a grand challenge in computational chemistry due to large sizes of constituent molecules and complex intra- and intermolecular interactions. While generative modeling has revolutionized structure discovery for molecules, inorganic solids, and metal-organic frameworks, extending such approaches to fully periodic molecular crystals is still elusive. Here, we present MolCrystalFlow, a flow-based generative model for molecular crystal structure prediction. The framework disentangles intramolecular complexity from intermolecular packing by embedding molecules as rigid bodies and jointly learning the lattice matrix, molecular orientations, and centroid positions. Centroids and orientations are represented on their native Riemannian manifolds, allowing geodesic flow construction and graph neural network operations that respects geometric symmetries. We benchmark our model against a state-of-the-art generative model (MOFFlow) for large-size periodic crystals and a rule-based structure generation method (Genarris) on two open-source molecular crystal datasets. MolCrystalFlow outperforms MOFFlow while achieving competitive performance against Genarris. We also demonstrate an integration of MolCrystalFlow model with universal machine learning potential to accelerate molecular crystal structure prediction, paving the way for data-driven generative discovery of molecular crystals.

cs.LG↗

arXiv2025

All that structure matches does not glitter

Generative models for materials, especially inorganic crystals, hold potential to transform the theoretical prediction of novel compounds and structures. Advancement in this field depends on robust benchmarks and minimal, information-rich datasets that enable meaningful model evaluation. This paper critically examines common datasets and reported metrics for a crystal structure prediction task$\unicode{x2014}$generating the most likely structures given the chemical composition of a material. We focus on three key issues: First, materials datasets should contain unique crystal structures; for example, we show that the widely-utilized carbon-24 dataset only contains $\approx$40% unique structures. Second, materials datasets should not be split randomly if polymorphs of many different compositions are numerous, which we find to be the case for the perov-5 and MP-20 datasets. Third, benchmarks can mislead if used uncritically, e.g., reporting a match rate metric without considering the structural variety exhibited by identical building blocks. To address these oft-overlooked issues, we introduce several fixes. We provide revised versions of the carbon-24 dataset: one with duplicates removed, one deduplicated and split by number of atoms $N$, one with enantiomorphs, and two containing only identical structures but with different unit cells. We also propose new splits for datasets with polymorphs, ensuring that polymorphs are grouped within each split subset, setting a more sensible standard for benchmarking model performance. Finally, we present METRe and cRMSE, new model evaluation metrics that can correct existing issues with the match rate metric.

cs.LG↗

arXiv2025

Open Materials Generation with Stochastic Interpolants

The discovery of new materials is essential for enabling technological advancements. Computational approaches for predicting novel materials must effectively learn the manifold of stable crystal structures within an infinite design space. We introduce Open Materials Generation (OMatG), a unifying framework for the generative design and discovery of inorganic crystalline materials. OMatG employs stochastic interpolants (SI) to bridge an arbitrary base distribution to the target distribution of inorganic crystals via a broad class of tunable stochastic processes, encompassing both diffusion models and flow matching as special cases. In this work, we adapt the SI framework by integrating an equivariant graph representation of crystal structures and extending it to account for periodic boundary conditions in unit cell representations. Additionally, we couple the SI flow over spatial coordinates and lattice vectors with discrete flow matching for atomic species. We benchmark OMatG's performance on two tasks: Crystal Structure Prediction (CSP) for specified compositions, and 'de novo' generation (DNG) aimed at discovering stable, novel, and unique structures. In our ground-up implementation of OMatG, we refine and extend both CSP and DNG metrics compared to previous works. OMatG establishes a new state of the art in generative modeling for materials discovery, outperforming purely flow-based and diffusion-based implementations. These results underscore the importance of designing flexible deep learning frameworks to accelerate progress in materials science. The OMatG code is available at https://github.com/FERMat-ML/OMatG.

cs.LG↗

arXiv2024

The emergence of bulk structure in clusters via isotropic multi-well pair potentials

The mechanical, optical, and chemical properties of a wide variety of soft materials are enabled and constrained by their bulk structure. How this structure emerges at small system sizes during self-assembly has been the subject of decades of research, with the aim of designing and controlling material functionality. Despite these efforts, it is still not fundamentally understood how nontrivial interparticle interactions in a finite $N$-body system influence resultant structure, and how that structure depends on $N$. In this study, we investigate the emergence of non-close packings using multi-well isotropic pair potentials to simulate finite cluster formation of four distinct two-dimensional crystal structures. These pair potentials encode multiple preferred length scales into the system, allowing us to understand how anisotropic structural motifs -- as opposed to close-packing -- emerge as cluster size $N$ increases. We find a tendency toward close-packing at small system sizes irrespective of the bulk structure; however, the system size at which bulk structure emerges is influenced by the coordination number of the bulk and the shape of the pair potential. Anisotropic structure emerges through the formation of bonds at a secondary bonding length at larger system sizes, and it is also dependent upon the shape of the pair potential. Our findings demonstrate that tuning particle-particle interactions can enable the engineering of nano- or mesoscale soft matter clusters, in applications as diverse as drug delivery and hierarchical materials design.

cond-mat.soft↗

arXiv2024

A snapshot review on soft-materials assembly design utilizing machine learning methods

Since the surge of data in materials science research and the advancement in machine learning methods, an increasing number of researchers are introducing machine learning techniques into the next generation of materials discovery, ranging from neural-network learned potentials to automated characterization techniques for experimental images. In this snapshot review, we first summarize the landscape of techniques for soft materials assembly design that do not employ machine learning or artificial intelligence and then discuss specific machine-learning and artificial-intelligence-based methods that enhance the design pipeline, such as high-throughput crystal-structure characterization and the inverse design of building blocks for materials assembly and properties. Additionally, we survey the landscape of current developments of scientific software, especially in the context of their compatibility with traditional molecular dynamics engines such as LAMMPS and HOOMD-blue.

cond-mat.soft↗

arXiv2024

Local structural features elucidate crystallization of complex structures

Complex crystal structures are composed of multiple local environments, and how this type of order emerges spontaneously during crystal growth has yet to be fully understood. We study crystal growth across various structures and along different crystallization pathways, using self-assembly simulations of identical particles that interact via multi-well isotropic pair potentials. We apply an unsupervised machine learning method to features from bond-orientational order metrics to identify different local motifs present during a given structure's crystallization process. In this manner, we distinguish different crystallographic sites in highly complex structures. Tailoring this newly developed order parameter to structures of varying complexity and coordination number, we study the emergence of local order along a multi-step crystal growth pathway -- from a low-density fluid to a high-density, supercooled amorphous liquid droplet and to a bulk crystal. We find a consistent under-coordination of the liquid relative to the average coordination number in the bulk crystal. We use our order parameter to analyze the geometrically frustrated growth of a Frank--Kasper phase and discover how structural defects compete with the formation of crystallographic sites that are higher-coordinated than the liquid environments. The here-presented method for classifying order on a particle-by-particle level have broad applicability to future studies of structural self-assembly and crystal growth, and they can aid in the design of building blocks and for targeting pathways of formation of novel soft-matter structures.

cond-mat.soft↗