Federated Hyperparameter Tuning - Challenges, Baselines, and Connections

This course explores the challenges and connections in federated hyperparameter tuning, focusing on adapting standard approaches to form baselines for the federated setting. The teaching method includes theoretical discussions and empirical demonstrations, showcasing a new method called FedEx that accelerates federated hyperparameter tuning. The intended audience for this course includes individuals interested in machine learning, federated learning, and hyperparameter optimization. Upon completion, learners will gain insights into the challenges of hyperparameter tuning in federated settings, learn about the FedEx method, and understand how it can outperform natural baselines in various benchmarks.