From Optimality to Robustness: Adaptive Re-Sampling Strategies in Stochastic Bandits

Dorian Baudry, Patrick Saux, Odalric-Ambrym Maillard

Keywords: Multi-armed bandits, Stochastic Bandits, robustness, regret analysis

TL;DR: We generalize a bandit algorithm that is optimal for bounded distributions to light-tailed unbounded distributions and obtain robust guarantees and strong practical performance under various hypotheses.

Abstract: The stochastic multi-arm bandit problem has been extensively studied under standard assumptions on the arm's distribution (e.g bounded with known support, exponential family, etc). These assumptions are suitable for many real-world problems but sometimes they require knowledge (on tails for instance) that may not be precisely accessible to the practitioner, raising the question of the robustness of bandit algorithms to model misspecification. In this paper we study a generic \emph{Dirichlet Sampling} (DS) algorithm, based on pairwise comparisons of empirical indices computed with \textit{re-sampling} of the arms' observations and a data-dependent \textit{exploration bonus}. We show that different variants of this strategy achieve provably optimal regret guarantees when the distributions are bounded and logarithmic regret for semi-bounded distributions with a mild quantile condition. We also show that a simple tuning achieve robustness with respect to a large class of unbounded distributions, at the cost of slightly worse than logarithmic asymptotic regret. We finally provide numerical experiments showing the merits of DS in a decision-making problem on synthetic agriculture data.

Code Of Conduct: I certify that all co-authors of this work have read and commit to adhering to the NeurIPS Statement on Ethics, Fairness, Inclusivity, and Code of Conduct.

Code: https://github.com/DBaudry/Dirichlet_Sampling_for_Bandits_Neurips21