Skip to content

FedCS

FedCS aims to optimize the number of clients involved in each round. It does so by greedily selecting the largest number of clients that can feasibly complete the round of training. This reduces the number of clients that participate, but fail to contribute to the global model in each round.

It is based on the paper

Client Selection for Federated Learning with Heterogeneous Resources in Mobile Edge.
Nishio, Takayuki, and Ryo Yonetani. 2018. arXiv [cs.NI]. arXiv. https://arxiv.org/abs/1804.08333.
The algorithm may be selected by choosing FedCS as the algorithm in the config file.

It requires the following parameters:

Key Description Example Value
pre_sampling Describes the percentage of the total available client to select for participation in client selection. If set to 0 disable preselection. 0.2
fixed_client_no Whether the client number should be fixed or not (useful for comparison) True or False
c Describes the share of all clients to participate in the round (only if fixed_client_no is True) 0.2