Chaotic-AES Integration for Privacy-Preserving Distributed Personalized Optimization
IEEE Signal Processing Letters · 2026 · 33 · 2854-2858
DOI: 10.1109/lsp.2026.3711542 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Qingfeng Tang und 4 weiteren Autor:innen. Die Arbeit nennt 47 Literaturangaben. Gefördert von: National Natural Science Foundation of China, NSF of Chongqing.
Abstract (Originalsprache)
This letter proposes CAES-PerDSGT, a lightweight privacy-preserving framework for personalized distributed stochastic gradient tracking (PerDSGT) in multi-agent networks. By integrating chaotic key variation with the Advanced Encryption Standard (AES) at the communication layer, the proposed method secures message exchange while preserving the integrity of the underlying optimization updates. Consequently, CAES-PerDSGT can be integrated into existing PerDSGT implementations with negligible computational and communication overhead. Numerical evaluations on personalized learning tasks with heterogeneous and non-IID data demonstrate that the algorithm protects under the stated threat models against both honest-but-curious internal agents and external eavesdroppers without compromising convergence stability.
Förderung
- National Natural Science Foundation of China
- NSF of Chongqing
Zitieren
APA
Tang, Q., Zhou, X., Yuan, Y., Lu, Q., & Li, H. (2026). Chaotic-AES Integration for Privacy-Preserving Distributed Personalized Optimization. IEEE Signal Processing Letters, 33, 2854-2858. https://doi.org/10.1109/lsp.2026.3711542
Vancouver
Tang Q, Zhou X, Yuan Y, Lu Q, Li H. Chaotic-AES Integration for Privacy-Preserving Distributed Personalized Optimization. IEEE Signal Processing Letters. 2026;33:2854-2858. doi:10.1109/lsp.2026.3711542
BibTeX
@article{Tang2026,
author = {Tang, Qingfeng and Zhou, Xiaohui and Yuan, Yi and Lu, Qingguo and Li, Huaqing},
title = {Chaotic-AES Integration for Privacy-Preserving Distributed Personalized Optimization},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {2854--2858},
doi = {10.1109/lsp.2026.3711542}
}
EUVIMED