Multi-Scale Zero-Order Optimization of Smooth Functions in an RKHS

TL;DR

This paper introduces a multi-scale UCB algorithm combining Gaussian process models with local polynomial estimators for optimizing smooth functions in RKHS, achieving near-optimal regret bounds and practical improvements.

Contribution

It proposes the LP-GP-UCB algorithm that integrates local polynomial estimators with GP models, providing new regret bounds for various kernels and practical advantages over existing methods.

Findings

Matches optimal performance for Squared Exponential kernel.

Achieves tighter regret bounds for Matérn kernels across all smoothness parameters.

Provides the first explicit regret bounds for Rational-Quadratic and Gamma-Exponential kernels.

Abstract

We aim to optimize a black-box function $f:\mathcal{X} \mapsto \mathbb{R}$ under the assumption that $f$ is H\"older smooth and has bounded norm in the RKHS associated with a given kernel $K$. This problem is known to have an agnostic Gaussian Process (GP) bandit interpretation in which an appropriately constructed GP surrogate model with kernel $K$ is used to obtain an upper confidence bound (UCB) algorithm. In this paper, we propose a new algorithm (\texttt{LP-GP-UCB}) where the usual GP surrogate model is augmented with Local Polynomial (LP) estimators of the H\"older smooth function $f$ to construct a multi-scale UCB guiding the search for the optimizer. We analyze this algorithm and derive high probability bounds on its simple and cumulative regret. We then prove that the elements of many common RKHS are H\"older smooth and obtain the corresponding H\"older smoothness parameters, and hence, specialize our regret bounds for several commonly used kernels. When specialized to the Squared Exponential (SE) kernel, \texttt{LP-GP-UCB} matches the optimal performance, while for the case of Mat\'ern kernels $(K_{\nu})_{\nu>0}$, it results in uniformly tighter regret bounds for all values of the smoothness parameter $\nu>0$. Most notably, for certain ranges of $\nu$, the algorithm achieves near-optimal bounds on simple and cumulative regrets, matching the algorithm-independent lower bounds up to polylog factors, and thus closing the large gap between the existing upper and lower bounds for these values of $\nu$. Additionally, our analysis provides the first explicit regret bounds, in terms of the budget $n$, for the Rational-Quadratic (RQ) and Gamma-Exponential (GE). Finally, experiments with synthetic functions as well as a CNN hyperparameter tuning task demonstrate the practical benefits of our multi-scale partitioning approach over some existing algorithms numerically.