ZeroHour
arXiv cs.AI / cs.LG / cs.CLpublished ()ingested Yilin Xu1

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

infoAI researchimportance 25
AI summary · glm-5.3

Paper releases a private client context once and confines adaptation to coefficients, matching full-model differential privacy with 2.67x less uplink on CIFAR-10.

The paper addresses the dimensionality misalignment between record-level differential privacy and low-dimensional client variation in personalized federated learning by releasing a private client context once and restricting repeated adaptation to a fixed coefficient space. A variable-length quantized Gaussian mechanism lets quantization error itself serve as the required privacy perturbation. On MNIST and CIFAR-10, the design matches or outperforms full-model private adaptation across privacy budgets and client heterogeneity while cutting protected uplink 2.67x at epsilon=16 on CIFAR-10.

  • Single private client-context release replaces repeated high-dimensional DP updates
  • Variable-length quantization error doubles as the Gaussian privacy noise
  • 2.67x protected uplink reduction at epsilon=16 on CIFAR-10 with comparable accuracy
  • Gains hold across privacy budgets and client heterogeneity on MNIST and CIFAR-10
Full article149 words · extracted from arxiv.org · click to collapse

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.

Text extracted automatically; images, tables and formatting may be missing. Original: https://arxiv.org/abs/2609.15950