PET-MAD, a universal interatomic potential for advanced materials modeling

2025-03-18

AI Summary

Existing machine learning interatomic potentials (MLIPs) exhibit only qualitative accuracy across the periodic table, suffer from strong bias toward low-energy configurations, and generalize poorly to molecules, organic materials, surfaces, and high-energy states. To address this, we propose PET-MAD—the first universal MLIP designed for materials spanning the entire periodic table—built upon the physics-informed PET architecture, trained on a high-consistency DFT dataset, and enhanced by systematic configuration sampling and atomic diversity augmentation. Its key contribution is the first unified modeling of inorganic solids, organic molecules, and surfaces, enabling near-quantitative prediction of thermal fluctuations, quantum effects, phase transitions, and functional properties. PET-MAD matches state-of-the-art specialized MLIPs across six material classes, enables out-of-the-box large-scale, long-timescale molecular dynamics simulations, and achieves full quantum accuracy with only minimal targeted retraining.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Multiagent Systems: Multiagent Learning
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Machine-learning interatomic potentials (MLIPs) have greatly extended the reach of atomic-scale simulations, offering the accuracy of first-principles calculations at a fraction of the effort. Leveraging large quantum mechanical databases and expressive architectures, recent"universal"models deliver qualitative accuracy across the periodic table but are often biased toward low-energy configurations. We introduce PET-MAD, a generally applicable MLIP trained on a dataset combining stable inorganic and organic solids, systematically modified to enhance atomic diversity. Using a moderate but highly-consistent level of electronic-structure theory, we assess PET-MAD's accuracy on established benchmarks and advanced simulations of six materials. PET-MAD rivals state-of-the-art MLIPs for inorganic solids, while also being reliable for molecules, organic materials, and surfaces. It is stable and fast, enabling, out-of-the-box, the near-quantitative study of thermal and quantum mechanical fluctuations, functional properties, and phase transitions. It can be efficiently fine-tuned to deliver full quantum mechanical accuracy with a minimal number of targeted calculations.

Problem

• Develops PET-MAD, a universal MLIP for diverse materials modeling.
• Addresses bias in MLIPs toward low-energy configurations.
• Enables accurate simulations of thermal, quantum, and phase transitions.

Innovation

• PET-MAD: universal MLIP for diverse materials
• Trained on stable inorganic and organic solids
• Efficiently fine-tuned for quantum mechanical accuracy