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

As a privacy-preserving method for implementing Vertical Federated Learning,Split Learning has been extensively researched. However, numerous studies haveindicated that the privacy-preserving capability of Split Learning isinsufficient. In this paper, we primarily focus on label inference attacks inSplit Learning under regression setting, which are mainly implemented throughthe gradient inversion method. To defend against label inference attacks, wepropose Random Label Extension (RLE), where labels are extended to obfuscatethe label information contained in the gradients, thereby preventing theattacker from utilizing gradients to train an attack model that can infer theoriginal labels. To further minimize the impact on the original task, wepropose Model-based adaptive Label Extension (MLE), where original labels arepreserved in the extended labels and dominate the training process. Theexperimental results show that compared to the basic defense methods, ourproposed defense methods can significantly reduce the attack model'sperformance while preserving the original task's performance.

Defending Label Inference Attacks in Split Learning under Regression Setting