Bayesian Optimization with Approximate Set Kernels
Abstract
We propose a practical Bayesian optimization method over sets, to minimize a black-box function that takes a set as a single input. Because set inputs are permutation-invariant, traditional Gaussian process-based Bayesian optimization strategies which assume vector inputs can fall short. To address this, we develop a Bayesian optimization method with \emph{set kernel} that is used to build surrogate functions. This kernel accumulates similarity over set elements to enforce permutation-invariance, but this comes at a greater computational cost. To reduce this burden, we propose two key components: (i) a more efficient approximate set kernel which is still positive-definite and is an unbiased estimator of the true set kernel with upper-bounded variance in terms of the number of subsamples, (ii) a constrained acquisition function optimization over sets, which uses symmetry of the feasible region that defines a set input. Finally, we present several numerical experiments which demonstrate that our method outperforms other methods.
Cite
@article{arxiv.1905.09780,
title = {Bayesian Optimization with Approximate Set Kernels},
author = {Jungtaek Kim and Michael McCourt and Tackgeun You and Saehoon Kim and Seungjin Choi},
journal= {arXiv preprint arXiv:1905.09780},
year = {2021}
}
Comments
18 pages, 7 figures, 5 tables, accepted for publication in Machine Learning Journal