Chapter 13: Hyper-parameter optimization in federated learning
Yi Zhou; Parikshit Ram; Theodoros Salonidis; Nathalie Baracaldo; Horst Samulowitz; Heiko Ludwig IBM Research, Yorktown Heights, NY, United States
Abstract
In this chapter, we study the relatively unexplored problem of hyper-parameter optimization (HPO) for federated learning (FL-HPO). We clearly formulate the FL-HPO problem, contrast it with traditional centralized HPO, and highlight the various potential novel challenges in the federated setup. Thoroughly walking through the HPO literature for federated learning, we discuss how existing FL-HPO approaches fare against the FL-HPO challenges. In particular, we introduce Federated Loss SuRface Aggregation (FLoRA), a general FL-HPO solution ...
Get Federated Learning now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
