ADAPT: Lightweight, Long-Range Machine Learning Force Fields Without Graphs

Year: 2025

Type: preprint

Abstract: Point defects play a central role in driving the properties of materials. First-principles methods are widely used to compute defect energetics and structures, including at scale for high-throughput defect databases. However, these methods are computationally expensive, making machine-learning force fields (MLFFs) an attractive alternative for accelerating structural relaxations. Most existing MLFFs are based on graph neural networks (GNNs), which can suffer from oversmoothing, oversquashing, and poor representation of long-range interactions. Both of these issues are especially of concern when modeling point defects. To address these challenges, we introduce the \textit{Accelerated Deep Atomic Potential Transformer} (ADAPT), an MLFF that replaces graph representations with a direct coordinates-in-space formulation and explicitly considers all pairwise atomic interactions. Atoms are treated as tokens, with a Transformer encoder modeling their interactions. Applied to a dataset of silicon point defects, ADAPT achieves a roughly 22% reduction in force and a roughly 40 percent reduction in energy prediction error relative to a state-of-the-art GNN-based model, while requiring only a fraction of the computational cost.

Source: arXiv (Cornell University)

DOI: 10.1103/3xf3-ts5f

Authors: Dramko, Evan, Xiong, Yihuang, Zhu, Yizhi, Hautier, Geoffroy, Reps, Thomas, Jermaine, Christopher, Kyrillidis, Anastasios

Institutions: Rice University, University of Wisconsin–Madison

Language: English

Topic: Machine Learning in Materials Science

Subfield: Materials Chemistry

Field: Materials Science

Domain: Physical Sciences

Open Access status: green

Funders: National Science Foundation, U.S. Department of Energy, Welch Foundation, Microsoft Research, Rice University, National Energy Research Scientific Computing Center, Office of Science, Basic Energy Sciences

OpenAlex ID: W7084166822