What Is Federated Learning?
In Chapter 1, Challenges in Big Data and Traditional AI, we examined how shifting tides in big data and machine learning (ML) have set the stage for a novel approach to practical ML applications. This chapter frames federated learning (FL) as the answer to the desire for this new ML approach. In a nutshell, FL is an approach to ML that allows models to be trained in parallel across data sources without the transmission of any data.
The goal of this chapter is to build up the case for the FL approach, with explanations of the necessary conceptual building blocks in order to ensure that you can achieve a similar understanding of the technical aspects and practical usage of FL.
After reading the chapter, you should have a high-level understanding of the FL process and should be able to visualize where the approach slots into real-world problem domains.
In this chapter, we will cover the following topics:
- Understanding the current state of ML...
