FedLab
Learn federated learning,
properly.
A dedicated platform for mastering federated learning — from the very fundamentals to research-level topics. Structured in weeks, with daily lessons and weekly final tests.
Built for wherever you are
Beginner
Complete beginners
No prior machine-learning experience assumed. Learn what federated learning is, why it exists, and how a basic federated system works.
Intermediate
Undergraduate CS/AI students
For students comfortable with basic ML. Core algorithms, optimization in federated settings, and hands-on system design.
Advanced
Graduate students
Privacy and security mechanisms, personalization, heterogeneity, and the mathematical foundations behind modern FL methods.
Start with Week 1
Foundations of Federated Learning- Day 1From Artificial Intelligence to Federated LearningDay 1 · The essential ideas behind learning from data, and why federated learning matters.
- Day 2The Federated Setting: Clients, Server, and RoundsThe anatomy of an FL system and the training loop.
- Day 3FedAvg: The Foundational AlgorithmFederated Averaging — the baseline every FL method is compared against.
- Day 4Data Heterogeneity: The Non-IID ChallengeWhy client data differs, and what that does to training.
- Day 5Privacy: Why Federated Learning Alone Is Not EnoughFL keeps data local — but model updates can still leak information.
- Day 6Communication and Systems ConstraintsThe engineering reality: bandwidth, stragglers, and dropouts.
- Day 7Week 1 Final TestTestCheck your understanding of the federated setting, FedAvg, heterogeneity, privacy, and systems constraints.