Adaptive Smooth Non-Stationary Bandits

TL;DR

This paper establishes the minimax dynamic regret rates for smooth non-stationary bandit models, develops adaptive algorithms without prior knowledge of smoothness parameters, and explores faster gap-dependent regret rates in environments with safe arms.

Contribution

It provides the first general minimax regret bounds for all smoothness parameters in non-stationary bandits and introduces adaptive algorithms that do not require prior parameter knowledge.

Findings

Established minimax dynamic regret rates for all parameters.

Designed adaptive algorithms achieving these rates without prior knowledge.

Identified conditions under which faster gap-dependent regret rates are possible.

Abstract

We study a $K$-armed non-stationary bandit model where rewards change smoothly, as captured by H\"{o}lder class assumptions on rewards as functions of time. Such smooth changes are parametrized by a H\"{o}lder exponent $\beta$ and coefficient $\lambda$. While various sub-cases of this general model have been studied in isolation, we first establish the minimax dynamic regret rate generally for all $K,\beta,\lambda$. Next, we show this optimal dynamic regret can be attained adaptively, without knowledge of $\beta,\lambda$. To contrast, even with parameter knowledge, upper bounds were only previously known for limited regimes $\beta\leq 1$ and $\beta=2$ (Slivkins, 2014; Krishnamurthy and Gopalan, 2021; Manegueu et al., 2021; Jia et al.,2023). Thus, our work resolves open questions raised by these disparate threads of the literature. We also study the problem of attaining faster gap-dependent regret rates in non-stationary bandits. While such rates are long known to be impossible in general (Garivier and Moulines, 2011), we show that environments admitting a safe arm (Suk and Kpotufe, 2022) allow for much faster rates than the worst-case scaling with $\sqrt{T}$. While previous works in this direction focused on attaining the usual logarithmic regret bounds, as summed over stationary periods, our new gap-dependent rates reveal new optimistic regimes of non-stationarity where even the logarithmic bounds are pessimistic. We show our new gap-dependent rate is tight and that its achievability (i.e., as made possible by a safe arm) has a surprisingly simple and clean characterization within the smooth H\"{o}lder class model.