Introducing Federated Learning for Internet of Things (IoT)

Federated learning (FL) refers to a distributed learning method that enables a number of internet of things devices to perform collaborative training of a particular machine learning model using their own local data. According to the network topology, the FL solutions can be categorized into three classes of systems, namely centralized FL, decentralized FL, and hierarchical FL. From a data‐partitioning perspective, the categorization of FL is performed according to the training data distributed according to the feature or sample spaces. In view of this, the FL can be taxonomized into three groups of methods, namely horizontal FL, vertical FL, and federated transfer learning. When it comes to the practical implementation of FL, there are a number of frameworks and software available. However, for proofs of concept and experimentations, open‐source frameworks are often enough. The chapter provides a general overview of the state‐of‐the‐art open‐source frameworks for FL.