On the Push-Based Asynchronous Federated Learning: A Bias-Correction Aggregation Approach
The article discusses a new framework called PushCen-ADFL for asynchronous decentralized federated learning. This framework aims to reduce communication overhead and improve model accuracy in heterogeneous systems. Experiments show that it can enhance accuracy by up to 6% while significantly lowering communication costs.
- ▪PushCen-ADFL is designed to enable stable training under asymmetric communication and delayed client participation.
- ▪The framework couples communication, aggregation, and local stabilization in a shared centroid representation space.
- ▪Experiments demonstrate that PushCen-ADFL improves accuracy under data heterogeneity by up to 6% and reduces per-push communication cost by more than 80%.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Machine Learning arXiv:2605.26162 (cs) [Submitted on 24 May 2026] Title:On the Push-Based Asynchronous Federated Learning: A Bias-Correction Aggregation Approach Authors:Jiahui Bai, Hai Dong, A. K. Qin View a PDF of the paper titled On the Push-Based Asynchronous Federated Learning: A Bias-Correction Aggregation Approach, by Jiahui Bai and 2 other authors View PDF HTML (experimental) Abstract:Asynchronous decentralized federated learning (ADFL) eliminates central coordination and global synchronization, making it attractive for large-scale and heterogeneous systems. However, frequent peer-to-peer communication, asynchronous updates on directed topologies, and non-IID data jointly lead to excessive communication overhead, biased aggregation and severe model drift.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
