Privacy-Aligned Personalized Federated Learning
Privacy-Aligned Personalized Federated Learning with Compact Adaptation and Variable-Length Gaussian Communication
This research provides a more efficient and privacy-preserving approach to federated learning, which is crucial for applications where data privacy is a top concern.
This paper introduces a new method for federated learning that addresses a misalignment issue in record-level differential privacy. It releases a private client context once and confines repeated adaptation to a fixed coefficient space, which reshapes noisy updates and reduces communication cost. Experiments on MNIST and CIFAR-10 show it matches or outperforms full-model private adaptation while reducing protected uplink by a factor of 2.67.
Privacy-Aligned Personalized Federated Learning with Compact Adaptation and Variable-Length Gaussian Communication
Record-level differential privacy exposes a structural misalignment in personalized federated learning when client-specific variation is low-dimensional while training repeatedly releases high-dimensional updates. In this paper, we address this misalignment by releasing a private client context once and confining repeated adaptation to a fixed coefficient space. Beyond dimensionality reduction, the factorized generator induces an adaptive optimization geometry that reshapes noisy updates, and controlled ablations show that most of its private-training gain is retained by radial evolution. To further reduce the communication cost, we realize the Gaussian mechanism for coefficient updates directly through variable-length quantization with finite expected code length, so that the quantization error itself serves as the required privacy perturbation rather than extra distortion. Across MNIST and CIFAR-10, our design matches or outperforms full-model private adaptation across privacy budgets and client heterogeneity, while reducing protected uplink by a factor of 2.67 at \(\varepsilon=16\) on CIFAR-10 with comparable future-client accuracy.