Horizontal Federated Learning

In this chapter, we introduce horizontal federated learning (HFL), covering the concept, architecture, application examples, and related works, as well as open research challenges.


HFL, a.k.a. sample-partitioned federated learning, or example-partitioned federated learning [Kairouz et al., 2019], can be applied in scenarios in which datasets at different sites share overlapping feature space but differ in sample space, as illustrated in Figure 4.1. It resembles the situation that data is horizontally partitioned inside a tabular view. In fact, the word “horizontal” comes from the term “horizontal partition,” which is widely used in the context of the traditional tabular view ...

Get Federated Learning now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.