mlsanityData to 14 Oct 2026

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

arXiv 2609.15950 ↗First posted 14 Sep 2026cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 9 of its 18 references that are in the map. First posted on arXiv 14 Sep 2026.

Builds on

9 of 18 references.

References · 9

  1. 2019Measuring the Effects of Non-Identical Data Distribution for Federated Visual Classificationcited by 851
  2. 2021Personalized Federated Learning using Hypernetworkscited by 212
  3. 2021Large Scale Private Learning via Low-rank Reparametrizationcited by 88
  4. 2023PeFLL: Personalized Federated Learning by Learning to Learncited by 21
  5. 2023Compression with Exact Error Distribution for Federated Learningcited by 11
  6. 2023Communication Efficient Private Federated Learning Using Ditheringcited by 9
  7. 2024Rejection-Sampled Universal Quantization for Smaller Quantization Errorscited by 4
  8. 2025Deploying Models to Non-participating Clients in Federated Learning without Fine-tuning: A Hypernetwork-based Approachcited by 2
  9. 2025Private Model Personalization Revisitedcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Enhanced Privacy and Communication Efficiency in Non-IID Federated Learning with Adaptive Quantization and Differential Privacy
  2. 2025Privacy-Preserving Decentralized Federated Learning via Explainable Adaptive Differential Privacy
  3. 2025Communication-Efficient and Privacy-Adaptable Mechanism for Federated Learning
  4. 2026Communication-Efficient and Privacy-Adaptable Mechanism -- a Federated Learning Scheme with Convergence Analysis
  5. 2024The Effect of Quantization in Federated Learning: A R'enyi Differential Privacy Perspective
  6. 2025Privacy-Preserving Quantized Federated Learning with Diverse Precision
  7. 2024CorBin-FL: A Differentially Private Federated Learning Mechanism using Common Randomness
  8. 2025FedRand: Enhancing Privacy in Federated Learning with Randomized LoRA Subparameter Updates