OMat24 releases a new open dataset of 110M+ DFT calculations and EquiformerV2 models achieving SOTA on Matbench Discovery with F1>0.9 for stability and 20 meV/atom accuracy for formation energies.
hub Mixed citations
A foundation model for atomistic materials chemistry
Mixed citation behavior. Most common role is method (44%).
34
Pith papers citing it
Method
44% of classified citations
abstract
Atomistic simulations of matter, especially those that leverage first-principles (ab initio) electronic structure theory, provide a microscopic view of the world, underpinning much of our understanding of chemistry and materials science. Over the last decade or so, machine-learned force fields have transformed atomistic modeling by enabling simulations of ab initio quality over unprecedented time and length scales. However, early ML force fields have largely been limited by: (i) the substantial computational and human effort of developing and validating potentials for each particular system of interest; and (ii) a general lack of transferability from one chemical system to the next. Here we show that it is possible to create a general-purpose atomistic ML model, trained on a public dataset of moderate size, that is capable of running stable molecular dynamics for a wide range of molecules and materials. We demonstrate the power of the MACE-MP-0 model - and its qualitative and at times quantitative accuracy - on a diverse set of problems in the physical sciences, including properties of solids, liquids, gases, chemical reactions, interfaces and even the dynamics of a small protein. The model can be applied out of the box as a starting or "foundation" model for any atomistic system of interest and, when desired, can be fine-tuned on just a handful of application-specific data points to reach ab initio accuracy. Establishing that a stable force-field model can cover almost all materials changes atomistic modeling in a fundamental way: experienced users get reliable results much faster, and beginners face a lower barrier to entry. Foundation models thus represent a step towards democratising the revolution in atomic-scale modeling that has been brought about by ML force fields.
hub tools
citation-role summary
method 4
background 2
baseline 2
dataset 1
citation-polarity summary
representative citing papers
Presents an end-to-end constraint-aware quantum optimization pipeline using XY-mixer QAOA and Grover Adaptive Search for low-energy defect configurations in doped ZrO2, with QAOA validated against exact enumeration on a high-accuracy QUBO surrogate of MACE energies.
ALMs unify pretrained atomistic encoder, LLM, and denoising diffusion via continuous projectors and staged training to reach SOTA on text-conditioned crystal prediction and de novo generation.
DPA4 is a new SE(3)-equivariant interatomic potential with EMFA SO(2) convolution that sets new accuracy-cost records on Matbench Discovery and SPICE benchmarks using fewer parameters than prior models.
JanusPipe introduces SymFold and WaveK to enable efficient 3D-parallel training for conservative MLIPs, reporting 1.51x and 1.45x average throughput gains over 1F1B and Hanayo baselines on 32 GPUs.
Lang2MLIP is an LLM multi-agent framework that automates end-to-end development of machine learning interatomic potentials from natural language input for heterogeneous materials systems.
A protocol is introduced to derive effective inertial and viscous-damping constants for nonequilibrium polarization dynamics in soft-mode ferroelectric PbTiO3.
Kernels from pretrained MLIP latent spaces outperform standard acquisition methods in active learning for reactive chemistry, reducing required labels by 38% for energy error and 28% for force error.
Torched-TACAW enables efficient large-scale STEM-EELS simulations of vibrational and magnon excitations in defective materials by combining ML-driven molecular dynamics with supercell partitioning and on-the-fly multislice processing.
Pre-registered validation of an ML Na-cathode voltage screen yields 0.67 V MAE against experiment, with Materials Project PBE+U references 0.54 V low and dominating the error.
MLIPilot deploys LLM agents to autonomously optimize MACE MLIP training on molecular and periodic datasets by proposing code edits and validating against a domain-specific scorecard.
Force-aware Neural Tangent Kernels combined with chunked acquisition provide scalable and distribution-robust active learning for MLIPs, outperforming baselines on OC20 and remaining competitive on other benchmarks.
Machine learning models, especially certain deep neural networks, can predict lattice thermal conductivity with useful accuracy across different generalization tests while being orders of magnitude faster than first-principles calculations.
Structural pruning of SO(3) equivariant atomistic models from large checkpoints yields 1.5-4x fewer parameters and 2.5-4x less pre-training compute than small models trained from scratch, while outperforming them on most Matbench Discovery metrics and downstream tasks.
VibroML automates remediation of dynamic instabilities in crystalline materials by combining MLIPs with genetic algorithms for polymorph search, finite-temperature MD validation, and compositional alloying to yield stable structures from databases like Alexandria.
Long-range ML potentials improve liquid silica but remain insufficient for experimental medium-range order in the glass, as shown by diffraction data and structural analyses.
Bayesian E(3)-equivariant MLPs with joint energy-force NLL loss achieve competitive accuracy while enabling uncertainty-guided active learning, OOD detection, and calibration.
Machine learning interatomic potentials fine-tuned on first-principles relaxation data accurately reproduce phonon spectra and optical lineshapes for defects, matching explicit calculations and experiments.
DAO pretrains Siamese diffusion-based models on stable/unstable crystal data to achieve 100% experimental match on Cr6Os2 and 2000x speedup over DFT on real superconductors.
MatterSim delivers a single deep learning force field that simulates inorganic materials across elements, 0-5000 K, and up to 1000 GPa with near first-principles accuracy for lattice dynamics, mechanics, and Gibbs free energies.
Statistical simulations show that local structural variations produce a continuous distribution of core-level binding energies that fully accounts for the observed XPS broadening in silicon oxides, matching layer-resolved sputtering data.
Quantum zero-point effects contribute less to low-temperature dislocation glide in bcc metals than earlier empirical-potential studies indicated, leaving the simulation-experiment discrepancy unresolved.
Systematic tests show naive fine-tuning excels for single-task accuracy while multihead replay best preserves out-of-distribution robustness in MLIP adaptation.
Split Ga vacancies are the most stable vacancy type in β-, α-, and κ-Ga2O3, identified via MLIP screening and confirmed with HSE06 DFT, limiting n-type doping unless grown oxygen-poor.
citing papers explorer
-
Constraint-Aware Quantum Optimization of Defect Configurations in Doped ZrO2: XY-Mixer QAOA and Grover Adaptive Search
Presents an end-to-end constraint-aware quantum optimization pipeline using XY-mixer QAOA and Grover Adaptive Search for low-energy defect configurations in doped ZrO2, with QAOA validated against exact enumeration on a high-accuracy QUBO surrogate of MACE energies.
-
Atomistic Language Models Understand and Generate Materials
ALMs unify pretrained atomistic encoder, LLM, and denoising diffusion via continuous projectors and staged training to reach SOTA on text-conditioned crystal prediction and de novo generation.
-
DPA4: Pushing the Accuracy-Cost Frontier of Interatomic Potentials with EMFA SO(2) Convolution
DPA4 is a new SE(3)-equivariant interatomic potential with EMFA SO(2) convolution that sets new accuracy-cost records on Matbench Discovery and SPICE benchmarks using fewer parameters than prior models.
-
JanusPipe: Efficient Pipeline Parallel Training for Machine Learning Interatomic Potentials
JanusPipe introduces SymFold and WaveK to enable efficient 3D-parallel training for conservative MLIPs, reporting 1.51x and 1.45x average throughput gains over 1F1B and Hanayo baselines on 32 GPUs.
-
Lang2MLIP: End-to-End Language-to-Machine Learning Interatomic Potential Development with Autonomous Agentic Workflows
Lang2MLIP is an LLM multi-agent framework that automates end-to-end development of machine learning interatomic potentials from natural language input for heterogeneous materials systems.
-
Effective dynamic constants for nonequilibrium third-principles simulations
A protocol is introduced to derive effective inertial and viscous-damping constants for nonequilibrium polarization dynamics in soft-mode ferroelectric PbTiO3.
-
Pretrained Model Representations as Acquisition Signals for Active Learning of MLIPs
Kernels from pretrained MLIP latent spaces outperform standard acquisition methods in active learning for reactive chemistry, reducing required labels by 38% for energy error and 28% for force error.
-
Efficient Large-Scale STEM-EELS Simulations With Torched-TACAW
Torched-TACAW enables efficient large-scale STEM-EELS simulations of vibrational and magnon excitations in defective materials by combining ML-driven molecular dynamics with supercell partitioning and on-the-fly multislice processing.
-
Computational references are not experiments: pre-registered validation of machine-learned sodium-cathode voltages
Pre-registered validation of an ML Na-cathode voltage screen yields 0.67 V MAE against experiment, with Materials Project PBE+U references 0.54 V low and dominating the error.
-
MLIPilot: LLM-Driven Auto-Research for Machine-Learned Interatomic Potentials
MLIPilot deploys LLM agents to autonomously optimize MACE MLIP training on molecular and periodic datasets by proposing code edits and validating against a domain-specific scorecard.
-
Force-Aware Neural Tangent Kernels for Scalable and Robust Active Learning of MLIPs
Force-aware Neural Tangent Kernels combined with chunked acquisition provide scalable and distribution-robust active learning for MLIPs, outperforming baselines on OC20 and remaining competitive on other benchmarks.
-
Fast and Accurate Prediction of Lattice Thermal Conductivity via Machine Learning Surrogates
Machine learning models, especially certain deep neural networks, can predict lattice thermal conductivity with useful accuracy across different generalization tests while being orders of magnitude faster than first-principles calculations.
-
Compact SO(3) Equivariant Atomistic Foundation Models via Structural Pruning
Structural pruning of SO(3) equivariant atomistic models from large checkpoints yields 1.5-4x fewer parameters and 2.5-4x less pre-training compute than small models trained from scratch, while outperforming them on most Matbench Discovery metrics and downstream tasks.
-
VibroML: an automated toolkit for high-throughput vibrational analysis and dynamic instability remediation of crystalline materials using machine-learned potentials
VibroML automates remediation of dynamic instabilities in crystalline materials by combining MLIPs with genetic algorithms for polymorph search, finite-temperature MD validation, and compositional alloying to yield stable structures from databases like Alexandria.
-
Neutron and X-ray Diffraction Reveal the Limits of Long-Range Machine Learning Potentials for Medium-Range Order in Silica Glass
Long-range ML potentials improve liquid silica but remain insufficient for experimental medium-range order in the glass, as shown by diffraction data and structural analyses.
-
Role of Local Structural Variation in X-ray Photoelectron Spectrum of Silicon Oxide Interfaces
Statistical simulations show that local structural variations produce a continuous distribution of core-level binding energies that fully accounts for the observed XPS broadening in silicon oxides, matching layer-resolved sputtering data.
-
Revisiting quantum effects on dislocation glide in bcc metals from DFT calculations and machine-learning potentials
Quantum zero-point effects contribute less to low-temperature dislocation glide in bcc metals than earlier empirical-potential studies indicated, leaving the simulation-experiment discrepancy unresolved.
-
Fine-tuning MLIP foundation models: strategies for accuracy and transferability
Systematic tests show naive fine-tuning excels for single-task accuracy while multihead replay best preserves out-of-distribution robustness in MLIP adaptation.
-
Universal Stability of Ga Split Vacancies across \alpha-, \beta-, and \kappa-Ga2O3 Polymorphs: A Machine-Learning Accelerated Study
Split Ga vacancies are the most stable vacancy type in β-, α-, and κ-Ga2O3, identified via MLIP screening and confirmed with HSE06 DFT, limiting n-type doping unless grown oxygen-poor.
-
Systematic Fine-Tuning of MACE Interatomic Potentials for Catalysis
Fine-tuned MACE MLIPs achieve lower mean absolute errors on catalytic reaction energies and barriers than from-scratch models, with a large fine-tuned model performing best on both metallic and oxide systems including out-of-distribution cases.
-
MatterSim-MT: A multi-task foundation model for in silico materials characterization
MatterSim-MT is a multi-task ML foundation model pretrained on 35M+ structures for in silico materials property prediction and complex simulations.
-
Exploring Multi-Transition-Metal NASICON Frameworks as High-Performance Cathodes for Sodium-Ion Batteries
DFT screening of unary to ternary NASICONs with Mn, Cr, Fe identifies Na_x MnFe0.5Cr0.5(PO4)3 as a candidate cathode balancing phase stability, ~4 V voltages from Fe redox, and 0.3-0.4 eV Na+ barriers.
-
Machine Learning Interatomic Potentials: Advancing Open-Source Software for Efficient and Scalable Molecular Simulation
mlip v2 is a new software release that integrates API redesign, e3j backend, eSEN model, improved charge modeling, and expanded simulation capabilities to support larger-scale molecular modeling.
-
Accurate and Efficient Interatomic Potentials for Dislocations in InP
New ACE and MACE potentials for InP achieve at most 4% error on partial dislocation formation energies versus DFT, outperforming literature models by factors of 4-12 while being computationally faster.
-
SLUSCHI-UP: A Web Infrastructure for SLUSCHI Melting-Temperature Calculations Using Universal Machine-Learning Interatomic Potentials
SLUSCHI-UP deploys the SLUSCHI melting-temperature workflow as a web service backed by selectable universal ML interatomic potentials with reported validation errors on benchmark sets.
-
From Knowledge to Action: Outcomes of the 2025 Large Language Model (LLM) Hackathon for Applications in Materials Science and Chemistry
Hackathon submissions indicate LLMs are moving from general assistants toward composable multi-agent systems for structuring scientific knowledge and automating tasks in materials science and chemistry.