Introducing Existing Federated Learning Frameworks
The objective of this chapter is to introduce existing federated learning (FL) frameworks and platforms, applying each to federated learning scenarios involving toy machine learning (ML) problems. The platforms focused on in this chapter are Flower, TensorFlow Federated, OpenFL, IBM FL, and STADLE – the idea behind this selection was to help you by covering a breadth of existing FL platforms.
By the end of this chapter, you should have a basic understanding of how to use each platform for FL, and you should be able to choose a platform based on its associated strengths and weaknesses for an FL application.
In this chapter, we will cover the following topics:
- Introduction to existing FL frameworks
- Implementations of an example NLP FL task on movie review dataset, using existing frameworks
- Implementations of example computer vision FL task with non-IID datasets, using existing frameworks
