Reproduce “LLM-Assisted Multi-Fidelity Bayesian Optimization: Fusing Expert Knowledge with Numerical Data” 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
Dongjun Han, Mingyu Lee, Ikjin Lee
The authors introduce an optimization method that uses expert-informed large language model comparisons to guide battery design when accurate simulations are costly. A pairwise Gaussian process turns judgments about which design needs a more positive residual correction—the difference left after global scaling—into a smooth basis, calibrated with accurate simulation data. On one battery simulation task, normalized regret AUC fell 61.2% versus conventional MFBO, showing faster progress in evaluations, not a better final optimum or universal gains.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.