Federated learning playground

Abstract

Federated learning is the leading paradigm for the collaborative training of machine learning models in a privacy-preserving manner. It was first introduced by Google as a learning paradigm for the next- word prediction model in google keyboards. In this project, students will build an interactive platform that supports experimentation of federated learning algorithms. It will serve as an educational visualization platform that provides analytical insights into the federated training process. This project may involve frameworks such as TensorFlow.js, D3.js, React Native etc. Some knowledge of python or tensorflow is recommended.