A Materials Foundation Model via Hybrid Invariant-Equivariant Architectures

2025-02-25

arXiv.org

AI Summary

Machine learning interatomic potentials (MLIPs) face a fundamental trade-off: invariant architectures offer computational efficiency but suffer from insufficient accuracy for higher-order physical quantities (e.g., forces, stresses), whereas equivariant architectures ensure high physical fidelity at the cost of substantial computational overhead. This work introduces HIENet, a foundation model for materials science, which—uniquely—provides a provably SE(3)-symmetric and physically constrained hybrid invariant-equivariant message-passing architecture. It integrates SE(3)-invariant graph neural networks with equivariant tensor propagation, augmented by physics-guided inter-layer coupling and symmetry-aware regularization. Evaluated on standard benchmarks including MD17 and Materials Project, HIENet achieves state-of-the-art accuracy across energy, force, and stress predictions. Moreover, it attains a 3.2× speedup in inference over the best-performing equivariant models, enabling accelerated downstream applications such as alloy screening and crystal structure optimization.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Humans and AI: Human-in-the-loop Machine Learning
• Computer Vision: Other Foundations of Computer Vision

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Machine learning interatomic potentials (MLIPs) can predict energy, force, and stress of materials and enable a wide range of downstream discovery tasks. A key design choice in MLIPs involves the trade-off between invariant and equivariant architectures. Invariant models offer computational efficiency but may not perform as well, especially when predicting high-order outputs. In contrast, equivariant models can capture high-order symmetries, but are computationally expensive. In this work, we propose HIENet, a hybrid invariant-equivariant materials interatomic potential model that integrates both invariant and equivariant message passing layers, while provably satisfying key physical constraints. HIENet achieves state-of-the-art performance with considerable computational speedups over prior models. Experimental results on both common benchmarks and downstream materials discovery tasks demonstrate the efficiency and effectiveness of HIENet.

Problem

• Balancing invariant and equivariant architectures in MLIPs
• Improving computational efficiency while capturing high-order symmetries
• Developing a hybrid model for accurate materials property prediction

Innovation

• Hybrid invariant-equivariant message passing layers
• Provably satisfies physical constraints
• State-of-the-art performance with speedups