Reproduce “Bayesian Optimization with Unknown Search Space” 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
Huong Ha, Santu Rana, Sunil Gupta, Thanh Nguyen, Hung Tran-The, Svetha Venkatesh
GPUCB-UBO uses Bayesian optimization to search beyond an initial bounded region for black-box problems. It uses a Gaussian-process model and an ε-accuracy checkpoint to decide when and how far to expand. On an MNIST multilayer perceptron hyperparameter task, it reaches 97.8% test-set accuracy after 12 iterations, while comparison methods require more than 24. This is an iteration-efficiency result, not a training-time result.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.