About this document

Regret Analysis of Improved GP-UCB by Old Cool is a document available to read on EtoBox.

This paper analyzes an improved version of the Gaussian Process Upper Confidence Bound (GP-UCB) algorithm for Bayesian optimization, called Improved Randomized GP-UCB (IRGP-UCB), which addresses the issue of increasing confidence parameters. The authors demonstrate that IRGP-UCB achieves sub-linear expected regret without the need for increasing the confidence parameter, particularly in finite input domains. Additionally, the paper includes numerical experiments to validate the proposed method

Author

Old Cool

Language

EN