UCB, wherein agents cooperatively minimize the group regret under the coordination of a central server.

By combining a caching algorithm, namely Greedy Dual Size with Frequency (GDSF) or Least Expected Cost (LEC), with a model multiplexer, we achieve optimal rates in both offline and online settings.

We fully resolve the conjecture posed in Subramanian et al. (2022),

We study the limits and capability of public-data assisted differentially private (P A-DP) algorithms. Specifically, we focus on the problem of stochastic convex optimization (SCO) with either labeled or unlabeled public data.

We consider distributed convex optimization problems in the regime when the communication between the server and the workers is expensive in both uplink and downlink directions.