Federated Learning (FL) enables collaborative model training across decentralized devices while keeping raw data local. The central server aggregates model updates (typically weights or gradients) using FedAvg algorithm.
Key Uses:
Healthcare (training diagnostic models across hospitals) and mobile keyboard prediction without uploading user text.
Hacks:
Implement differential privacy by adding Gaussian noise to gradients. Use secure multi-party computation (SMPC) or homomorphic encryption for model aggregation. Apply model compression (sparsification) before transmission to reduce bandwidth. Hierarchical FL for cross-silo and cross-device scenarios.
References:
McMahan, B., et al. (2017). Communication-Efficient Learning of Deep Networks from Decentralized Data. AISTATS.