Chapter 10: Graph-aware federated learning
Songtao Lua; Pengwei Xingb; Han Yub aIBM Thomas J. Watson Research Center, Yorktown Heights, NY, United StatesbNanyang Technological University, Singapore, Singapore
Abstract
As data owners distributedly sampled their data locally, federated learning (FL) often needs to deal with data heterogeneity. In this chapter, we investigate FL scenarios in which the distributions of the local datasets are non-i.i.d. Existing graph-aware FL (GFL) approaches can be classified into: 1) decentralized optimization-based FL, 2) multi-centered FL, and 3) graph-aided FL. This chapter introduces these GFL paradigms with the mathematical problem formulation, and then presents corresponding algorithms for solving these problems, ...
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.
