Interactive lesson~18 minAdvanced
Federated Learning
Federated learning trains models across many devices without centralizing raw data. The server aggregates updates, not private examples.
PrivacyAggregationDifferential privacy
Mental model
Bring the model to the data instead of the data to the model.
Privacy-sensitive healthcare, keyboard prediction, finance, and edge AI can benefit from decentralized learning.
Privacy
balanced71% modeled signal
Accuracy
balanced58% modeled signal
Fairness
balanced55% modeled signal
Concept pipeline
Build the idea in four moves
Interactive lab
Train across edge devices without collecting raw data.
Broadcast
Send the current model to clients.
Focus lens
The part that usually clicks late
Non-IID data
Each client may represent a different distribution.
Privacy
71
Accuracy
58
Fairness
55
Knowledge check
What is federated learning’s core promise?
Next horizon
Where this topic is headed
Federated analytics
Personalized FL
Private fine-tuning