Bayesian Optimization with Exponential Convergence

TL;DR

This paper introduces a Bayesian optimization technique that guarantees exponential convergence without auxiliary optimization or delta-cover sampling, making it more practical and efficient.

Contribution

It proposes a novel Bayesian optimization method that removes the need for auxiliary optimization and delta-cover sampling, achieving exponential convergence.

Findings

Achieves exponential convergence rate in Bayesian optimization

Eliminates auxiliary optimization and delta-cover sampling requirements

Simplifies implementation and improves practicality

Abstract

This paper presents a Bayesian optimization method with exponential convergence without the need of auxiliary optimization and without the delta-cover sampling. Most Bayesian optimization methods require auxiliary optimization: an additional non-convex global optimization problem, which can be time-consuming and hard to implement in practice. Also, the existing Bayesian optimization method with exponential convergence requires access to the delta-cover sampling, which was considered to be impractical. Our approach eliminates both requirements and achieves an exponential convergence rate.