Algorithmic Differentiation for Automated Modeling of Machine Learned Force Fields
This addresses a key bottleneck for the force field community by enabling more versatile and efficient modeling, though it is incremental in automating existing processes.
The paper tackles the challenge of efficiently reconstructing force fields from atomistic simulation data by proposing the use of algorithmic differentiation in machine learning modeling, which automates the inclusion of novel descriptors and improves computational efficiency by an order of magnitude.
Reconstructing force fields (FFs) from atomistic simulation data is a challenge since accurate data can be highly expensive. Here, machine learning (ML) models can help to be data economic as they can be successfully constrained using the underlying symmetry and conservation laws of physics. However, so far, every descriptor newly proposed for an ML model has required a cumbersome and mathematically tedious remodeling. We therefore propose using modern techniques from algorithmic differentiation within the ML modeling process -- effectively enabling the usage of novel descriptors or models fully automatically at an order of magnitude higher computational efficiency. This paradigmatic approach enables not only a versatile usage of novel representations and the efficient computation of larger systems -- all of high value to the FF community -- but also the simple inclusion of further physical knowledge such as higher-order information (e.g. Hessians, more complex partial differential equations constraints etc.), even beyond the presented FF domain.