Best-of-Both-Worlds Linear Contextual Bandits
Abstract: This study investigates the problem of $K$-armed linear contextual bandits, an instance of the multi-armed bandit problem, under an adversarial corruption. At each round, a decision-maker observes an independent and identically distributed context and then selects an arm based on the context and past observations. After selecting an arm, the decision-maker incurs a loss corresponding to the selected arm. The decision-maker aims to minimize the cumulative loss over the trial. The goal of this study is to develop a strategy that is effective in both stochastic and adversarial environments, with theoretical guarantees. We first formulate the problem by introducing a novel setting of bandits with adversarial corruption, referred to as the contextual adversarial regime with a self-bounding constraint. We assume linear models for the relationship between the loss and the context. Then, we propose a strategy that extends the {\tt RealLinExp3} by \citet{Neu2020} and the Follow-The-Regularized-Leader (FTRL). The regret of our proposed algorithm is shown to be upper-bounded by $O\left(\min\left\{\frac{(\log(T))^3}{\Delta_{*}} + \sqrt{\frac{C(\log(T))^3}{\Delta_{*}}},\ \ \sqrt{T}(\log(T))^2\right\}\right)$, where $T \in\mathbb{N}$ is the number of rounds, $\Delta_{*} > 0$ is the constant minimum gap between the best and suboptimal arms for any context, and $C\in[0, T] $ is an adversarial corruption parameter. This regret upper bound implies $O\left(\frac{(\log(T))^3}{\Delta_{*}}\right)$ in a stochastic environment and by $O\left( \sqrt{T}(\log(T))^2\right)$ in an adversarial environment. We refer to our strategy as the {\tt Best-of-Both-Worlds (BoBW) RealFTRL}, due to its theoretical guarantees in both stochastic and adversarial regimes.
Submission Length: Regular submission (no more than 12 pages of main content)
Previous TMLR Submission Url: https://openreview.net/forum?id=PR9P35ZkOW&nesting=2&sort=date-desc
Changes Since Last Submission: In the first post on December 18th, the 'times' package had inadvertently been included in the Tex file, which caused the font to change. In this post, that 'times' package has been removed.
Assigned Action Editor: ~Chicheng_Zhang1
Submission Number: 1986
Loading