English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Federated learning (FL) is a burgeoning distributed machine learningframework where a central parameter server (PS) coordinates many local users totrain a globally consistent model. Conventional federated learning inevitablyrelies on a centralized topology with a PS. As a result, it will paralyze oncethe PS fails. To alleviate such a single point failure, especially on the PS,some existing work has provided decentralized FL (DFL) implementations likeCDSGD and D-PSGD to facilitate FL in a decentralized topology. However, thereare still some problems with these methods, e.g., significant divergencebetween users' final models in CDSGD and a network-wide model average necessityin D-PSGD. In order to solve these deficiency, this paper devises a new DFLimplementation coined as DACFL, where each user trains its model using its owntraining data and exchanges the intermediate models with its neighbors througha symmetric and doubly stochastic matrix. The DACFL treats the progress of eachuser's local training as a discrete-time process and employs a first orderdynamic average consensus (FODAC) method to track the \textit{average model} inthe absence of the PS. In this paper, we also provide a theoretical convergenceanalysis of DACFL on the premise of i.i.d data to strengthen its rationality.The experimental results on MNIST, Fashion-MNIST and CIFAR-10 validate thefeasibility of our solution in both time-invariant and time-varying networktopologies, and declare that DACFL outperforms D-PSGD and CDSGD in most cases.

DACFL: Dynamic Average Consensus Based Federated Learning in Decentralized Topology