Reproduce “Batch Bayesian Optimization on Permutations using the Acquisition Weighted Kernel” 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
Changyong Oh, Roberto Bondesan, Efstratios Gavves, Max Welling
LAW2ORDER selects diverse, promising batches for expensive permutation optimization. It weights a determinantal point process kernel by acquisition values, then greedily fills each batch with high-variance candidates and evaluates them in parallel. On TSP-burma14, LAW-EST reports cost 3369 ± 7 versus 3786 ± 74 for unweighted DPP-MAX-EST, with batch size 5 over 15 runs. LAW-EST is competitive across problems, but other methods lead on some benchmarks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.