KL-UCB-switch: optimal regret bounds for stochastic bandits from both a distribution-dependent and a distribution-free viewpoints

Abstract

We consider $K$-armed stochastic bandits and consider cumulative regret bounds up to time $T$. We are interested in strategies achieving simultaneously a distribution-free regret bound of optimal order $\sqrt{KT}$ and a distribution-dependent regret that is asymptotically optimal, that is, matching the $\kappa \ln T$ lower bound by Lai and Robbins (1985) and Burnetas and Katehakis (1996), where $\kappa$ is the optimal problem-dependent constant. This constant $\kappa$ depends on the model $\mathcal{D}$ considered (the family of possible distributions over the arms). Ménard and Garivier (2017) provided strategies achieving such a bi-optimality in the parametric case of models given by one-dimensional exponential families, while Lattimore (2016, 2018) did so for the family of (sub)Gaussian distributions with variance less than $1$. We extend this result to the non-parametric case of all distributions over $[0,1]$. We do so by combining the MOSS strategy by Audibert and Bubeck (2009), which enjoys a distribution-free regret bound of optimal order $\sqrt{KT}$, and the KL-UCB strategy by Cappé et al. (2013), for which we provide in passing the first analysis of an optimal distribution-dependent $\kappa \ln T$ regret bound in the model of all distributions over $[0,1]$. We were able to obtain this non-parametric bi-optimality result while working hard to streamline the proofs (of previously known regret bounds and thus of the new analyses carried out); a second merit of the present contribution is therefore to provide a review of proofs of classical regret bounds for index-based strategies for $K$-armed stochastic bandits.

References in corpus (1)

• A Finite-Time Analysis of Multi-armed Bandits Problems with Kullback-Leibler Divergences

Cited by in corpus (9)

• Nearly Minimax-Optimal Regret for Linearly Parameterized Bandits
• Efficient Change-Point Detection for Tackling Piecewise-Stationary Bandits
• On the bias, risk and consistency of sample means in multi-armed bandits
• Polynomial Cost of Adaptation for X -Armed Bandits
• Corralling Stochastic Bandit Algorithms
• Finite-time Analysis of Globally Nonstationary Multi-Armed Bandits
• On Regret with Multiple Best Arms
• Accelerated learning from recommender systems using multi-armed bandit
• Ballooning Multi-Armed Bandits