Deep Ensembles vs. Committees for Uncertainty Estimation in Neural-Network Force Fields: Comparison and Application to Active Learning

TL;DR

This paper compares deep ensembles, committees, and bootstrap methods for uncertainty estimation in neural-network force fields, demonstrating improved active learning workflows for materials modeling.

Contribution

It introduces a generalized deep-ensemble method with multiheaded neural networks and heteroscedastic loss, enhancing uncertainty estimation in force fields.

Findings

Deep ensembles outperform committees in uncertainty estimation.

Active learning significantly reduces training data needs.

Fast training enables practical active learning workflows.

Abstract

A reliable uncertainty estimator is a key ingredient in the successful use of machine-learning force fields for predictive calculations. Important considerations are correlation with error, overhead during training and inference, and efficient workflows to systematically improve the force field. However, in the case of neural-network force fields, simple committees are often the only option considered due to their easy implementation. Here we present a generalization of the deep-ensemble design, based on multiheaded neural networks and a heteroscedastic loss, that can efficiently deal with uncertainties in both the energy and the forces. We compare uncertainty metrics based on deep ensembles, committees and bootstrap-aggregation ensembles using data for an ionic liquid and a perovskite surface. We demonstrate an adversarial approach to active learning to efficiently and progressively refine the force fields. That active learning workflow is realistically possible thanks to exceptionally fast training enabled by residual learning and a nonlinear learned optimizer.