Multi-Message Shuffled Privacy in Federated Learning
This work addresses privacy-preserving federated learning for distributed systems with communication limitations, representing a specific technical advancement rather than a broad breakthrough.
The paper tackles the problem of differentially private distributed optimization under communication constraints by developing a communication-efficient private distributed mean estimation scheme using the multi-message shuffled privacy framework, achieving order-optimal privacy-communication-performance tradeoffs and resolving open questions about shuffled models and private vector sums.
We study differentially private distributed optimization under communication constraints. A server using SGD for optimization aggregates the client-side local gradients for model updates using distributed mean estimation (DME). We develop a communication-efficient private DME, using the recently developed multi-message shuffled (MMS) privacy framework. We analyze our proposed DME scheme to show that it achieves the order-optimal privacy-communication-performance tradeoff resolving an open question in [1], whether the shuffled models can improve the tradeoff obtained in Secure Aggregation. This also resolves an open question on the optimal trade-off for private vector sum in the MMS model. We achieve it through a novel privacy mechanism that non-uniformly allocates privacy at different resolutions of the local gradient vectors. These results are directly applied to give guarantees on private distributed learning algorithms using this for private gradient aggregation iteratively. We also numerically evaluate the private DME algorithms.