Learn2Hop: Learned Optimization on Rough Landscapes

TL;DR

This paper introduces Learn2Hop, a meta-learned optimizer that efficiently explores complex non-convex landscapes with many local minima, significantly improving low-energy configuration discovery in atomic systems.

Contribution

The work presents a novel learned optimization algorithm that adapts to rough landscapes, demonstrating improved exploration and generalization in atomic structural optimization tasks.

Findings

Learn2Hop exhibits a 'hopping' behavior for efficient exploration.

The optimizer accelerates discovery of low energy minima.

It generalizes well to unseen atomic configurations.

Abstract

Optimization of non-convex loss surfaces containing many local minima remains a critical problem in a variety of domains, including operations research, informatics, and material design. Yet, current techniques either require extremely high iteration counts or a large number of random restarts for good performance. In this work, we propose adapting recent developments in meta-learning to these many-minima problems by learning the optimization algorithm for various loss landscapes. We focus on problems from atomic structural optimization--finding low energy configurations of many-atom systems--including widely studied models such as bimetallic clusters and disordered silicon. We find that our optimizer learns a 'hopping' behavior which enables efficient exploration and improves the rate of low energy minima discovery. Finally, our learned optimizers show promising generalization with efficiency gains on never before seen tasks (e.g. new elements or compositions). Code will be made available shortly.