Advancing AI for materials with MatterSim: experimental synthesis, faster simulation, and multi-task models

2026-05-12 · Source: Microsoft Research · Field: Science & Research — Artificial Intelligence & Machine Learning, Engineering & Applied Sciences, Mathematics & Computational Sciences · Depth: Expert, long

Summary

Microsoft Research has released significant updates to its MatterSim platform, a deep learning atomistic model for materials design. Key advancements include the experimental validation of MatterSim-v1's prediction of tetragonal tantalum phosphorus (TaP) as a high-performance thermal conductor, with a measured thermal conductivity of 152 W/m/K, comparable to silicon. The MatterSim-v1 model also received performance improvements, achieving 3-5x faster inference speeds and integration with the LAMMPS software for multi-GPU simulations. Additionally, a new multi-task foundation model, MatterSim-MT, has been introduced. This model is pretrained on over 35 million first-principles-labeled structures and can predict energies, forces, stress, and various material properties, enabling complex simulations like vibrational spectroscopy, ferroelectric switching, and electrochemical redox processes.

Key takeaway

For materials scientists and engineers focused on novel materials discovery, these MatterSim updates significantly accelerate the design cycle. You can now leverage experimentally validated predictions for thermal conductors and utilize MatterSim-MT for comprehensive, multi-property simulations, moving from computational screening to targeted experimental follow-up more efficiently. Consider integrating the faster MatterSim-v1 with LAMMPS for large-scale, multi-GPU workflows.

Key insights

MatterSim advancements enable faster, more accurate, and multi-property materials simulations, validated by experimental discovery.

Principles

• ML interatomic potentials accelerate materials design.
• High-throughput screening narrows experimental search space.

Method

MatterSim-MT uses a multi-task architecture pretrained on 35M+ structures, fine-tuned for properties like Bader charges, to simulate complex material phenomena beyond potential energy surfaces.

In practice

• Use MatterSim-v1 for high-throughput thermal conductor screening.
• Integrate MatterSim-v1 with LAMMPS for scaled simulations.
• Apply MatterSim-MT for multi-property material characterization.

Topics

• MatterSim
• MatterSim-MT
• Materials Characterization
• Thermal Conductivity
• Machine Learning Interatomic Potentials

Code references

• microsoft/mattersim

Best for: AI Scientist, Research Scientist, Machine Learning Engineer

Related on AIssential

• Inverse design of bespoke interatomic potentials via active learning by information-matching — The information-matching (IM) approach tailors interatomic potentials by linking training data selection to specific material property uncertainty targets.
• AI materials discovery now needs to move into the real world — AI-driven autonomous labs aim to revolutionize materials discovery by accelerating synthesis and testing, despite current limitations.
• Testing LLMs on superconductivity research questions — Curated data sources significantly enhance LLM accuracy and reliability in specialized scientific domains.
• LEAP: A closed-loop framework for perovskite precursor additive discovery — Coupling a domain-specialized LLM with active learning and expert review accelerates perovskite additive discovery under low-data conditions.
• AI Models Trained on Physics Are Changing Engineering — Large physics models drastically accelerate engineering design by replacing or augmenting traditional physics simulations.
• Coupling Language Models with Physics-based Simulation for Synthesis of Inorganic Materials — Coupling LLMs with physics-based simulations offers more viable inorganic material synthesis routes than classical algorithms, leveraging LLMs' implicit priors.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Microsoft Research.