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

Data valuation aims to quantify the usefulness of individual data sources intraining machine learning (ML) models, and is a critical aspect of data-centricML research. However, data valuation faces significant yet frequentlyoverlooked privacy challenges despite its importance. This paper studies thesechallenges with a focus on KNN-Shapley, one of the most practical datavaluation methods nowadays. We first emphasize the inherent privacy risks ofKNN-Shapley, and demonstrate the significant technical difficulties in adaptingKNN-Shapley to accommodate differential privacy (DP). To overcome thesechallenges, we introduce TKNN-Shapley, a refined variant of KNN-Shapley that isprivacy-friendly, allowing for straightforward modifications to incorporate DPguarantee (DP-TKNN-Shapley). We show that DP-TKNN-Shapley has severaladvantages and offers a superior privacy-utility tradeoff compared to naivelyprivatized KNN-Shapley in discerning data quality. Moreover, even non-privateTKNN-Shapley achieves comparable performance as KNN-Shapley. Overall, ourfindings suggest that TKNN-Shapley is a promising alternative to KNN-Shapley,particularly for real-world applications involving sensitive data.

Threshold KNN-Shapley: A Linear-Time and Privacy-Friendly Approach to Data Valuation