A universal augmentation framework for long-range electrostatics in machine learning interatomic potentials

2025-07-18

AI Summary

Most machine-learned interatomic potentials (MLIPs) rely on short-range approximations and thus fail to accurately capture long-range electrostatic interactions. To address this, we propose Latent Ewald Summation (LES), a general, plug-and-play framework for implicit long-range electrostatic enhancement. LES jointly infers Born effective charges, atomic polarizabilities, and electrostatic interactions solely from energy and force labels—without explicit training on electronic properties. It implements a differentiable, Ewald-based implicit learning module, ensuring native compatibility with state-of-the-art MLIPs including MACE, NequIP, CACE, and CHGNet. Experiments demonstrate that LES substantially improves prediction accuracy for complex systems—including liquid water, polar dipeptides, and gold dimers/defective substrates. On the SPICE benchmark, MACE-LES-OFF not only outperforms its short-range counterpart but also faithfully reproduces molecular dipole moments and Born effective charges. LES establishes a new paradigm for universal, first-principles–accurate MLIP construction.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Multiagent Systems: Multiagent Learning

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Most current machine learning interatomic potentials (MLIPs) rely on short-range approximations, without explicit treatment of long-range electrostatics. To address this, we recently developed the Latent Ewald Summation (LES) method, which infers electrostatic interactions, polarization, and Born effective charges (BECs), just by learning from energy and force training data. Here, we present LES as a standalone library, compatible with any short-range MLIP, and demonstrate its integration with methods such as MACE, NequIP, CACE, and CHGNet. We benchmark LES-enhanced models on distinct systems, including bulk water, polar dipeptides, and gold dimer adsorption on defective substrates, and show that LES not only captures correct electrostatics but also improves accuracy. Additionally, we scale LES to large and chemically diverse data by training MACELES-OFF on the SPICE set containing molecules and clusters, making a universal MLIP with electrostatics for organic systems including biomolecules. MACELES-OFF is more accurate than its short-range counterpart (MACE-OFF) trained on the same dataset, predicts dipoles and BECs reliably, and has better descriptions of bulk liquids. By enabling efficient long-range electrostatics without directly training on electrical properties, LES paves the way for electrostatic foundation MLIPs.

Problem

• Enabling long-range electrostatics in ML interatomic potentials
• Integrating LES with diverse short-range MLIP methods
• Improving accuracy in organic and biomolecular systems

Innovation

• Latent Ewald Summation for long-range electrostatics
• Compatible with any short-range MLIP
• Scales to large chemically diverse datasets