Keynote: Federated Learning with Flower: AI's Next Frontier
summary
Claims that federated learning is usable from Python by shipping models instead of data and that Flower lets developers implement it in under 20 lines of code. Demo uses PyTorch, a Flower server and 3 simulated clients, runs FedAvg rounds and shows distributed loss falling and accuracy 14%→21% after a few rounds. Framework lists support for cross‑silo and cross‑device settings, differential privacy and secure aggregation, and the demo code actually runs end to end.
AI-written summary · about
chapters (10)
- 0:00Federated learning importance
- 1:08Data-rich vs data-poor divide
- 8:55AI training unsustainability
- 12:08Moving computation not data
- 16:44Federated Averaging explained
- 23:54Real-world FL examples
- 32:00Flower framework intro
- 38:48Live federated demo
- 48:07Future of federated AI
- 51:09Q&A on FL challenges
related talks
32mInnovation in the Age of Regulation: Federated Learning with Flower
32mPlant a Touch-Me-Not: Train Models Without Anyone Touching Your Data with Flower
42mLearning Together, Distributed: An Introduction to Federated Learning
01h 22mThe future of AI training is federated
18mEnabling Medical AI through Federated Learning
31mFrom Theory to Practice: Advances in Real-World Federated Learning
26mMachine Learning without the Data #federatedlearning
13mScalable and Portable Federated Learning Simulation Engine