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

Safeguarding the Intellectual Property (IP) of data has become criticallyimportant as machine learning applications continue to proliferate, and theirsuccess heavily relies on the quality of training data. While variousmechanisms exist to secure data during storage, transmission, and consumption,fewer studies have been developed to detect whether they are already leaked formodel training without authorization. This issue is particularly challengingdue to the absence of information and control over the training processconducted by potential attackers.  In this paper, we concentrate on the domain of tabular data and introduce anovel methodology, Local Distribution Shifting Synthesis (\textsc{LDSS}), todetect leaked data that are used to train classification models. The coreconcept behind \textsc{LDSS} involves injecting a small volume of syntheticdata--characterized by local shifts in class distribution--into the owner'sdataset. This enables the effective identification of models trained on leakeddata through model querying alone, as the synthetic data injection results in apronounced disparity in the predictions of models trained on leaked andmodified datasets. \textsc{LDSS} is \emph{model-oblivious} and hence compatiblewith a diverse range of classification models, such as Naive Bayes, DecisionTree, and Random Forest. We have conducted extensive experiments on seven typesof classification models across five real-world datasets. The comprehensiveresults affirm the reliability, robustness, fidelity, security, and efficiencyof \textsc{LDSS}.

From Zero to Hero: Detecting Leaked Data through Synthetic Data Injection and Model Querying