Adversarial Bandits with Multi-User Delayed Feedback: Theory and Application
This addresses the problem of resource allocation and online advertising for groups of users, extending prior single-user models, but it is incremental as it adapts existing methods to a multi-user setting.
The paper tackles the adversarial multi-armed bandit problem with delayed feedback from multiple users, proposing MUD-EXP3 and AMUD-EXP3 algorithms that achieve a regret bound of O(√(TM²ln N(N e + 4d_max))) and sublinear regret for unknown T, respectively.
The multi-armed bandit (MAB) models have attracted significant research attention due to their applicability and effectiveness in various real-world scenarios such as resource allocation, online advertising, and dynamic pricing. As an important branch, the adversarial MAB problems with delayed feedback have been proposed and studied by many researchers recently where a conceptual adversary strategically selects the reward distributions associated with each arm to challenge the learning algorithm and the agent experiences a delay between taking an action and receiving the corresponding reward feedback. However, the existing models restrict the feedback to be generated from only one user, which makes models inapplicable to the prevailing scenarios of multiple users (e.g. ad recommendation for a group of users). In this paper, we consider that the delayed feedback results are from multiple users and are unrestricted on internal distribution. In contrast, the feedback delay is arbitrary and unknown to the player in advance. Also, for different users in a round, the delays in feedback have no assumption of latent correlation. Thus, we formulate an adversarial MAB problem with multi-user delayed feedback and design a modified EXP3 algorithm MUD-EXP3, which makes a decision at each round by considering the importance-weighted estimator of the received feedback from different users. On the premise of known terminal round index $T$, the number of users $M$, the number of arms $N$, and upper bound of delay $d_{max}$, we prove a regret of $\mathcal{O}(\sqrt{TM^2\ln{N}(N\mathrm{e}+4d_{max})})$. Furthermore, for the more common case of unknown $T$, an adaptive algorithm AMUD-EXP3 is proposed with a sublinear regret with respect to $T$. Finally, extensive experiments are conducted to indicate the correctness and effectiveness of our algorithms.