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 (FL) is a popular paradigm for private and collaborativemodel training on the edge. In centralized FL, the parameters of a globalarchitecture (such as a deep neural network) are maintained and distributed bya central server/controller to clients who transmit model updates (gradients)back to the server based on local optimization. While many efforts have focusedon reducing the communication complexity of gradient transmission, the vastmajority of compression-based algorithms assume that each participating clientis able to download and train the current and full set of parameters, which maynot be a practical assumption depending on the resource constraints of smallerclients such as mobile devices. In this work, we propose a simple yet effectivenovel algorithm, Comfetch, which allows clients to train large networks usingreduced representations of the global architecture via the count sketch, whichreduces local computational and memory costs along with bi-directionalcommunication complexity. We provide a nonconvex convergence guarantee andexperimentally demonstrate that it is possible to learn large models, such as adeep convolutional network, through federated training on their sketchedcounterparts. The resulting global models exhibit competitive test accuracyover CIFAR10/100 classification when compared against un-compressed modeltraining.

Comfetch: Federated Learning of Large Networks on Constrained Clients via Sketching