Privacy-Aligned Personalized Federated Learning with Compact Adaptation and Variable-Length Gaussian Communication

arXiv:2609.15950 · cs.LG · Submitted 2026-09-14 · Read on arXiv

cs.LG

Submitted: 2026-09-14

Updated: 2026-09-14

Comments: 8 pages, 1 figure

License: http://creativecommons.org/licenses/by/4.0/

The gist: Record-level differential privacy exposes a structural misalignment in personalized federated learning when client-specific variation is low-dimensional while training repeatedly releases

Terminology

Abstract

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 epsilon=16 on CIFAR-10 with comparable future-client accuracy.

Related papers