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

Providing privacy protection has been one of the primary motivations ofFederated Learning (FL). Recently, there has been a line of work onincorporating the formal privacy notion of differential privacy with FL. Toguarantee the client-level differential privacy in FL algorithms, the clients'transmitted model updates have to be clipped before adding privacy noise. Suchclipping operation is substantially different from its counterpart of gradientclipping in the centralized differentially private SGD and has not beenwell-understood. In this paper, we first empirically demonstrate that theclipped FedAvg can perform surprisingly well even with substantial dataheterogeneity when training neural networks, which is partly because theclients' updates become similar for several popular deep architectures. Basedon this key observation, we provide the convergence analysis of a differentialprivate (DP) FedAvg algorithm and highlight the relationship between clippingbias and the distribution of the clients' updates. To the best of ourknowledge, this is the first work that rigorously investigates theoretical andempirical issues regarding the clipping operation in FL algorithms.

Understanding Clipping for Federated Learning: Convergence and Client-Level Differential Privacy