Federated UCBVI: Communication-Efficient Federated Regret Minimization with Heterogeneous Agents

Safwan Labbi, Daniil Tiapkin, Lorenzo Mancini +2

In this paper, we present the Federated Upper Confidence Bound Value Iteration algorithm ($\text{Fed-UCBVI}$), a novel extension of the $\text{UCBVI}$ algorithm (Azar et al., 2…