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

Top-$k$ frequent items detection is a fundamental task in data stream mining.Many promising solutions are proposed to improve memory efficiency while stillmaintaining high accuracy for detecting the Top-$k$ items. Despite the memoryefficiency concern, the users could suffer from privacy loss if participatingin the task without proper protection, since their contributed local datastreams may continually leak sensitive individual information. However, mostexisting works solely focus on addressing either the memory-efficiency problemor the privacy concerns but seldom jointly, which cannot achieve a satisfactorytradeoff between memory efficiency, privacy protection, and detection accuracy.  In this paper, we present a novel framework HG-LDP to achieve accurateTop-$k$ item detection at bounded memory expense, while providing rigorouslocal differential privacy (LDP) protection. Specifically, we identify two keychallenges naturally arising in the task, which reveal that directly applyingexisting LDP techniques will lead to an inferior ``accuracy-privacy-memoryefficiency'' tradeoff. Therefore, we instantiate three advanced schemes underthe framework by designing novel LDP randomization methods, which address thehurdles caused by the large size of the item domain and by the limited space ofthe memory. We conduct comprehensive experiments on both synthetic andreal-world datasets to show that the proposed advanced schemes achieve asuperior ``accuracy-privacy-memory efficiency'' tradeoff, saving $2300\times$memory over baseline methods when the item domain size is $41,270$. Our code isopen-sourced via the link.

Local Differentially Private Heavy Hitter Detection in Data Streams with Bounded Memory