Skip to content

Truncated automatic sparse differentiation for machine learning interatomic potentials

Sep 2026 · arXiv.org · Vol abs/2609.20510 · 4 citations · 52 references
Physics Computer Science

TL;DR

This work explains how to compute the sparsity pattern for MLIP derivatives and demonstrates that, for multiple foundation MLIPs, ASD computes full Hessians of large porous materials exactly, but with modest speedups at best.

Abstract

Machine learning interatomic potentials (MLIPs) learn the mapping from atomic positions to potential energy. The forces, the negative gradient of this energy, drive molecular dynamics and are readily obtained using automatic differentiation. Higher-order derivatives, most notably the Hessian, describe collective motion and allow the direct prediction of experimental observables, but are considered computationally inaccessible for large systems. We suggest a solution: in physical systems, interactions decay with distance, and most MLIPs build on this locality through message passing up to a finite receptive field. This implies both sparsity of higher-order derivatives and their decay with distance. This structure can be exploited using automatic sparse differentiation (ASD). We explain how to compute the sparsity pattern for MLIP derivatives and demonstrate that, for multiple foundation MLIPs, ASD computes full Hessians of large porous materials exactly, but with modest speedups at best. The larger gains come from truncated ASD: discarding small, but nonzero, Hessian entries between distant atoms yields order-of-magnitude speedups with negligible impact on predicted observables.

View source

Similar papers

Preprint Oct 2026

Overcoming the limitations of body-ordered potentials for atomistic machine learning

Machine-learning interatomic potentials have become indispensable tools in atomistic simulations. The most popular models rely heavily on physical priors, such as locality, smoothness and symmetry, that are assumed to improve the accuracy and transferability of the trained models. However, recent evidence suggests that...

Filippo Bigi, M. Domina, Maximilian L. Ach et al. · 1 citation
#machine learning Preprint Sep 2026

Hessian-based molecular conformation augmentation for a scalable and efficient strategy of machine learning interatomic potentials

While machine-learning interatomic potentials (MLIPs) have successfully learned potential energy surfaces (PES) and atomic forces, many practical applications, such as vibrational analysis and transition state search, rely heavily on the PES Hessian. Yet standard MLIPs are trained on energy and forces alone, and existi...

Bumju Kwak, Jeonghee Jo · 0 citations
Review Oct 2026

Specialized machine learning force fields for materials dynamics

Machine learning interatomic potentials (MLIPs) are transforming atomistic simulations by accessing unprecedented length and time scales. While pretrained equivariant graph neural networks achieve robust zero-shot performance for near-equilibrium properties across broad chemical spaces, their translation to complex mat...

Yue Wu, Hao-Yu Wan, Yuan Tian et al. · 0 citations
Open access Sep 2026

Symbolic ensemble learning enables discovery of fast accurate physics-based interatomic potentials

A hybrid symbolic-neural framework is introduced that unifies the interpretability of the Embedded Atom Method with the adaptability of data-driven learning, and derives a composite potential that surpasses the fidelity of its constituent models.

Bilvin Varughese, Aditya Koneru, Adil Muhammad et al. · 0 citations
#machine learning Preprint Oct 2026

BranchIP: Learning Adaptive Equivariant Computation for Interatomic Potentials

Equivariant machine learning interatomic potentials (MLIPs) have revolutionized atomistic modeling, but accurate treatment of complex materials and molecular systems demands expensive models. This limits simulation length- and time-scales, with tensor products a key computational bottleneck. The recent emergence of fou...

Laura Zichi, Gil Harari, Chuin Wei Tan et al. · 0 citations
Preprint Sep 2026

Multipole splats for optimized and inverted effective potentials

In modern density functional theory, effective Hamiltonians are constructed to reproduce densities and forces of electrons at equilibrium. However, their nonlocal potentials leave systematic errors in spectral properties and real-time dynamics. Access to local optimized or inverted effective potentials would remove thi...

Matija Medvidović, Angel Rubio, Juan Carrasquilla · 1 citation

Related blog posts

MIT News · Artificial Intelligence Nov 17, 2026

Learning the language of machine learning

By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.

Microsoft Research Blog Nov 16, 2026

TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient

AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.