Completed
Interlude: Follow-the-regularized-leader (FTRL)
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Federated Learning with Formal User-Level Differential Privacy Guarantees
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 (Non)-convex learning
- 3 Differential Privacy
- 4 Cross-device federated learning
- 5 Differentially private stochastic gradient descent DP
- 6 DP-SGD: Key insights
- 7 DP-Federated Averaging (DP-FedAvg)
- 8 Challenges for Amplification by Sampling in FL
- 9 Deconstructing the SGD model update
- 10 Noise Accumulation in Prefix Sums
- 11 Towards Tree Aggregation
- 12 Interlude: Follow-the-regularized-leader (FTRL)
- 13 DP-Follow-the-regularized leader (DP-FTRL)
- 14 DP-FTRL: Online learning properties
- 15 Privacy-Utility Trade-offs for Stackoverflow
- 16 Production model with formal DP
- 17 Matrix factorization view of prefix sum estimation
- 18 Matrix factorization view of DP prefix sum
- 19 Future directions
- 20 Acknowledgements
