Distributed Clustering of Linear Bandits in Peer to Peer Networks
This addresses efficient decision-making in distributed systems like sensor networks or collaborative AI, offering incremental improvements in communication-limited settings.
The paper tackles distributed linear bandit problems in peer-to-peer networks with limited communication, proposing two algorithms: one for identical problems achieving optimal asymptotic regret, and another for clustered problems that discovers clusters while maintaining optimal regret within each. Experiments on real-world datasets show performance compared to state-of-the-art.
We provide two distributed confidence ball algorithms for solving linear bandit problems in peer to peer networks with limited communication capabilities. For the first, we assume that all the peers are solving the same linear bandit problem, and prove that our algorithm achieves the optimal asymptotic regret rate of any centralised algorithm that can instantly communicate information between the peers. For the second, we assume that there are clusters of peers solving the same bandit problem within each cluster, and we prove that our algorithm discovers these clusters, while achieving the optimal asymptotic regret rate within each one. Through experiments on several real-world datasets, we demonstrate the performance of proposed algorithms compared to the state-of-the-art.