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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Motivated by high resource costs of centralized machine learning schemes aswell as data privacy concerns, federated learning (FL) emerged as an efficientalternative that relies on aggregating locally trained models rather thancollecting clients' potentially private data. In practice, available resourcesand data distributions vary from one client to another, creating an inherentsystem heterogeneity that leads to deterioration of the performance ofconventional FL algorithms. In this work, we present a federatedquantization-based self-supervised learning scheme (Fed-QSSL) designed toaddress heterogeneity in FL systems. At clients' side, to tackle dataheterogeneity we leverage distributed self-supervised learning while utilizinglow-bit quantization to satisfy constraints imposed by local infrastructure andlimited communication resources. At server's side, Fed-QSSL deploysde-quantization, weighted aggregation and re-quantization, ultimately creatingmodels personalized to both data distribution as well as specificinfrastructure of each client's device. We validated the proposed algorithm onreal world datasets, demonstrating its efficacy, and theoretically analyzedimpact of low-bit training on the convergence and robustness of the learnedmodels.

Fed-QSSL: A Framework for Personalized Federated Learning under Bitwidth and Data Heterogeneity