Kernel Methods for Cooperative Multi-Agent Contextual Bandits
This addresses cooperative learning in multi-agent systems with network delays, offering improved performance but is incremental as it builds on existing multi-agent bandit methods.
The paper tackles the kernelised contextual bandit problem for cooperative multi-agent decision-making, proposing Coop-KernelUCB, which achieves near-optimal per-agent regret bounds and outperforms existing benchmarks on synthetic and real-world network tests.
Cooperative multi-agent decision making involves a group of agents cooperatively solving learning problems while communicating over a network with delays. In this paper, we consider the kernelised contextual bandit problem, where the reward obtained by an agent is an arbitrary linear function of the contexts' images in the related reproducing kernel Hilbert space (RKHS), and a group of agents must cooperate to collectively solve their unique decision problems. For this problem, we propose \textsc{Coop-KernelUCB}, an algorithm that provides near-optimal bounds on the per-agent regret, and is both computationally and communicatively efficient. For special cases of the cooperative problem, we also provide variants of \textsc{Coop-KernelUCB} that provides optimal per-agent regret. In addition, our algorithm generalizes several existing results in the multi-agent bandit setting. Finally, on a series of both synthetic and real-world multi-agent network benchmarks, we demonstrate that our algorithm significantly outperforms existing benchmarks.