> ML_LITERATURE // BONAWITZ-2017-PRACTICAL-SECURE-AGGREGATION-FOR-PRIVACY-PRESERVING-MACHINE-LEARNING_v1.0
Practical Secure Aggregation for Privacy-Preserving Machine Learning
Keith Bonawitz, Vladimir Ivanov, Ben Kreuter, Antonio Marcedone, H. Brendan McMahan, Sarvar Patel, Daniel Ramage, Aaron Segal, Karn Seth · ACM SIGSAC Conference on Computer and Communications Security (CCS) (2017)
systems2017industry-standardartifactsAvailable
Principal Contribution
Invented a cryptographic multiparty protocol for Secure Aggregation (SecAgg), allowing a server to compute sum of model updates without inspecting any individual user update.
Operational Relevance
Serves as qualified reference for implementing task-federated-learning, task-privacy-preserving-ml in production systems.
Assumptions
- Underlying computational topology and mathematical bounds adhere to established convexity/smoothness guarantees
Limitations
- Hardware runtime speedups, privacy budgets, and convergence depend on hyperparameters and network communication limits
Connected Algorithms, Architectures & Tools
Related Algorithms:
Related Architectures:
Implementing Libraries: