Reproduce “Bayesian Model Discovery with Language Models: Closing the Loop between Hypotheses and Experiments” 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
JooKyung Yuh, Byungmin Park, Eunseo Choi, Jinheon Baek, Junmo Cho, Kyu Hyun Choi, Sung Ju Hwang, Kyong Hwan Jin and 1 more
Bayesian model discovery (BMD) combines language model proposals with numerical fitting and evidence-based model selection to choose experiments that distinguish candidate laws and refine parameters. Across four simulated benchmarks, BMD has the lowest prediction error among reported baselines. In a separate test with the true family supplied, error is 0.214 versus 1.000 for direct in-context learning; real-world performance remains untested.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.