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

We develop a new approach to tackle communication constraints in adistributed learning problem with a central server. We propose and analyze anew algorithm that performs bidirectional compression and achieves the sameconvergence rate as algorithms using only uplink (from the local workers to thecentral server) compression. To obtain this improvement, we design MCM, analgorithm such that the downlink compression only impacts local models, whilethe global model is preserved. As a result, and contrary to previous works, thegradients on local servers are computed on perturbed models. Consequently,convergence proofs are more challenging and require a precise control of thisperturbation. To ensure it, MCM additionally combines model compression with amemory mechanism. This analysis opens new doors, e.g. incorporating workerdependent randomized-models and partial participation.

Preserved central model for faster bidirectional compression in distributed settings