FedLess: Secure and Scalable Serverless Federated Learning
summary
FedLess is an open‑source, cloud‑agnostic FL framework that runs client code as serverless functions on AWS Lambda, Google Cloud Functions, Azure Functions and four other FaaS platforms, implemented in Python 3 with TensorFlow 2+/Keras and MongoDB as the parameter server. It uses local differential privacy (Gaussian noise), a FedScan semi‑asynchronous DBSCAN client‑selection and staleness‑aware aggregation (tau) to tolerate cold starts and stragglers. In tests with up to 200 concurrent clients it cuts training time ~8% on average, lowers cost ~20%, and gains ~2% final accuracy under heavy straggling.
AI-written summary · about
chapters (9)
- 0:00FedLess introduction
- 1:03Cloud computing evolution
- 4:52Federated Learning basics
- 7:03Serverless + Federated Learning
- 8:11FedLess architecture
- 13:25Addressing stragglers in Federated Learning
- 20:55Results and cost analysis
- 24:50Personalized models
- 28:03Q&A
related talks
31mFedLess: Secure and Scalable Serverless Federated Learning
42mFederated Learning: Unlocking the Values of Discrete Data for Enterprise
42mLearning Together, Distributed: An Introduction to Federated Learning
31mFrom Theory to Practice: Advances in Real-World Federated Learning
41mFederated Learning: Machine Learning on Decentralized Data
30mFederated Learning from simulation to production
31mFederated Machine Learning with Kubernetes
45mManaged federated learning on AWS: A case study for healthcare