Publications
CONFERENCE (INTERNATIONAL) Improved Regret Bounds for Gaussian Process Upper Confidence Bound in Bayesian Optimization
Shogo Iwazaki
The Thirty-ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025)
December 05, 2025
This paper addresses the Bayesian optimization problem (also referred to as the Bayesian setting of the Gaussian process bandit), where the learner seeks to minimize the regret under a function drawn from a known Gaussian process (GP). Under a Mat'ern kernel with some extent of smoothness, we show that the Gaussian process upper confidence bound (GP-UCB) algorithm achieves $\tilde{O}(\sqrt{T})$ cumulative regret with high probability. Furthermore, our analysis yields $O(\sqrt{T \ln^2 T})$ regret under a squared exponential kernel. These results fill the gap between the existing regret upper bound of GP-UCB and the current best upper bound provided by Scarlett [2018]. The key idea in our proof is to capture the concentration behavior of the input sequence realized by GP-UCB, enabling us to handle GP's information gain in a refined manner.
Paper :
Improved Regret Bounds for Gaussian Process Upper Confidence Bound in Bayesian Optimization
(external link)
PDF : Improved Regret Bounds for Gaussian Process Upper Confidence Bound in Bayesian Optimization