Part of Advances in Neural Information Processing Systems 34 (NeurIPS 2021)
Amin Karbasi, Vahab Mirrokni, Mohammad Shadravan
How can we make use of information parallelism in online decision-making problems while efficiently balancing the exploration-exploitation trade-off? In this paper, we introduce a batch Thompson Sampling framework for two canonical online decision-making problems with partial feedback, namely, stochastic multi-arm bandit and linear contextual bandit. Over a time horizon $T$, our batch Thompson Sampling policy achieves the same (asymptotic) regret bound of a fully sequential one while carrying out only $O(\log T)$ batch queries. To achieve this exponential reduction, i.e., reducing the number of interactions from $T$ to $O(\log T)$, our batch policy dynamically determines the duration of each batch in order to balance the exploration-exploitation trade-off. We also demonstrate experimentally that dynamic batch allocation outperforms natural baselines.