MACS, a multi-agent RL optimizer for periodic crystals, achieves faster local geometry optimization with fewer energy calls and a lower failure rate than BFGS, FIRE, and other baselines, and transfers zero-shot to chemically related unseen compositions.
Learning to optimize molecular geometries using reinforcement learning.Journal of Chemical Theory and Computation, 17(2):818–825, 2021
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
MACS: Multi-Agent Reinforcement Learning for Optimization of Crystal Structures
MACS, a multi-agent RL optimizer for periodic crystals, achieves faster local geometry optimization with fewer energy calls and a lower failure rate than BFGS, FIRE, and other baselines, and transfers zero-shot to chemically related unseen compositions.