Reproduce “Agent-Guided Discovery of Exchange-Correlation Functionals with Physics-based Validation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Achuth Chandrasekhar, Amir Barati Farimani
The authors use language-model agents to propose compact exchange–correlation formulas, which a numerical evaluator fits on training data and screens before separate held-out and self-consistent tests. Their 13-parameter candidate O159 has an 11.7% lower self-consistent test error than ωB97M-V on MGCDB84, but its external comparison is comparable, not a significant improvement, on 36 selected datasets.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.