Privatized Graph Federated Learning
This addresses privacy and efficiency problems for distributed machine learning systems, but appears incremental as it builds on existing federated learning with graph-based modifications.
The paper tackles the robustness and privacy issues in federated learning by introducing graph federated learning (GFL) with graph homomorphic perturbations for differential privacy, showing theoretical convergence and privacy analyses with performance illustrated through simulations.
Federated learning is a semi-distributed algorithm, where a server communicates with multiple dispersed clients to learn a global model. The federated architecture is not robust and is sensitive to communication and computational overloads due to its one-master multi-client structure. It can also be subject to privacy attacks targeting personal information on the communication links. In this work, we introduce graph federated learning (GFL), which consists of multiple federated units connected by a graph. We then show how graph homomorphic perturbations can be used to ensure the algorithm is differentially private. We conduct both convergence and privacy theoretical analyses and illustrate performance by means of computer simulations.