The future of AI training is federated
summary
Argues that production AI training moves to federated setups and shows a working Flower stack: flower new, flower run, flower list, flower log, flower stop; demo covers three runtimes — local simulation (10 clients by default, sampled 5 with fraction_fit=0.5, 3 server rounds in the demo, can scale to 100–200), remote Superlink server for hosted runs, and full deployment with Superlink/Supernode via Docker/Kubernetes to Raspberry Pi devices; uses Ray for simulation, live multi‑node demo succeeds albeit with occasional networking/flaky client failures.
AI-written summary · about
chapters (13)
- 0:00Federated learning intro
- 1:13Why federated AI is future
- 2:58Setting up Flower tutorial
- 6:12What is Flower
- 8:40Flower runtimes and deploy
- 10:01Building FL apps with Flower
- 13:34Local simulations with Flower
- 37:20Configuring FL parameters
- 39:07Simulations on remote superlinks
- 53:07Deploying Flower in systems
- 1:00:31Live demo with Raspberry Pis
- 1:09:21Next steps and ecosystem
- 1:17:30Q&A
related talks
57mKeynote: Federated Learning with Flower: AI's Next Frontier
32mInnovation in the Age of Regulation: Federated Learning with Flower
32mPlant a Touch-Me-Not: Train Models Without Anyone Touching Your Data with Flower
18mEnabling Medical AI through Federated Learning
13mScalable and Portable Federated Learning Simulation Engine
34mLFN AI Taskforce: Platform-Driven Fede...
31mFederated Machine Learning with Kubernetes- 31mFederated Machine Learning with Kubernetes