On Large-Cohort Training for Federated Learning

Zachary Charles, Zachary Garrett, Zhouyuan Huo, Sergei Shmulyian, Virginia Smith

Keywords: federated learning, distributed optimization, large batch training, large cohort training

TL;DR: We explore the impact of the cohort size (ie. how many clients participate in a round) on the optimization dynamics of federated learning.

Abstract: Federated learning methods typically learn a model by iteratively sampling updates from a population of clients. In this work, we explore how the number of clients sampled at each round (the cohort size) impacts the quality of the learned model and the training dynamics of federated learning algorithms. Our work poses three fundamental questions. First, what challenges arise when trying to scale federated learning to larger cohorts? Second, what parallels exist between cohort sizes in federated learning, and batch sizes in centralized learning? Last, how can we design federated learning methods that effectively utilize larger cohort sizes? We give partial answers to these questions based on extensive empirical evaluation. Our work highlights a number of challenges stemming from the use of larger cohorts. While some of these (such as generalization issues and diminishing returns) are analogs of large-batch training challenges, others (including catastrophic training failures and fairness concerns) are unique to federated learning.

Code Of Conduct: I certify that all co-authors of this work have read and commit to adhering to the NeurIPS Statement on Ethics, Fairness, Inclusivity, and Code of Conduct.

Code: https://github.com/google-research/federated/tree/f4e26c1b9b47ac320e520a8b9943ea2c5324b8c2/large_cohort

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/arxiv:2106.07820/code)