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

Machine learning (ML) classifiers are invaluable building blocks that havebeen used in many fields. High quality training dataset collected from multipledata providers is essential to train accurate classifiers. However, it raisesconcern about data privacy due to potential leakage of sensitive information intraining dataset. Existing studies have proposed many solutions toprivacy-preserving training of ML classifiers, but it remains a challengingtask to strike a balance among accuracy, computational efficiency, andsecurity. In this paper, we propose Heda, an efficient privacypreserving schemefor training ML classifiers. By combining homomorphic cryptosystem (HC) withdifferential privacy (DP), Heda obtains the tradeoffs between efficiency andaccuracy, and enables flexible switch among different tradeoffs by parametertuning. In order to make such combination efficient and feasible, we presentnovel designs based on both HC and DP: A library of building blocks based onpartially HC are proposed to construct complex training algorithms withoutintroducing a trusted thirdparty or computational relaxation; A set oftheoretical methods are proposed to determine appropriate privacy budget and toreduce sensitivity. Security analysis demonstrates that our solution canconstruct complex ML training algorithm securely. Extensive experimentalresults show the effectiveness and efficiency of the proposed scheme.

When Homomorphic Cryptosystem Meets Differential Privacy: Training Machine Learning Classifier with Privacy Protection