Federated Learning
Domain: MLOps
🎯 Motivation
Training models across multiple devices without collecting data centrally. Privacy-preserving.
📋 Concrete Example
Federated: 1000 phones each train on local data. Only model updates (not data) are sent to server. Server combines updates. Privacy preserved.
🔗 Analogy
A study group where each student studies at home, then shares what they learned (not their notes). The group combines insights. Federated learning trains across devices without sharing data.