Problem-Parameter-Free Federated Learning

Background and Motivation Problem Formulation of Federated Learning (FL) • Federated Learning:  Multiple clients collaboratively train a machine learning model with the help of a central server.  Each client performs multiple local update based on private data  Server aggregates the global model Advantages: • Ensures privacy by avoiding raw data sharing • Offers scalability and communication efficiency • Figure from https://encyclopedia.pub/entry/48625