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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Federated learning (FedL) has emerged as a popular technique for distributingmodel training over a set of wireless devices, via iterative local updates (atdevices) and global aggregations (at the server). In this paper, we developparallel successive learning (PSL), which expands the FedL architecture alongthree dimensions: (i) Network, allowing decentralized cooperation among thedevices via device-to-device (D2D) communications. (ii) Heterogeneity,interpreted at three levels: (ii-a) Learning: PSL considers heterogeneousnumber of stochastic gradient descent iterations with different mini-batchsizes at the devices; (ii-b) Data: PSL presumes a dynamic environment with dataarrival and departure, where the distributions of local datasets evolve overtime, captured via a new metric for model/concept drift. (ii-c) Device: PSLconsiders devices with different computation and communication capabilities.(iii) Proximity, where devices have different distances to each other and theaccess point. PSL considers the realistic scenario where global aggregationsare conducted with idle times in-between them for resource efficiencyimprovements, and incorporates data dispersion and model dispersion with localmodel condensation into FedL. Our analysis sheds light on the notion of coldvs. warmed up models, and model inertia in distributed machine learning. Wethen propose network-aware dynamic model tracking to optimize the modellearning vs. resource efficiency tradeoff, which we show is an NP-hardsignomial programming problem. We finally solve this problem through proposinga general optimization solver. Our numerical results reveal new findings on theinterdependencies between the idle times in-between the global aggregations,model/concept drift, and D2D cooperation configuration.

Parallel Successive Learning for Dynamic Distributed Model Training over Heterogeneous Wireless Networks