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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Building advanced machine learning (ML) models requires expert knowledge andmany trials to discover the best architecture and hyperparameter settings.Previous work demonstrates that model information can be leveraged to assistother attacks, such as membership inference, generating adversarial examples.Therefore, such information, e.g., hyperparameters, should be keptconfidential. It is well known that an adversary can leverage a target MLmodel's output to steal the model's information. In this paper, we discover anew side channel for model information stealing attacks, i.e., models'scientific plots which are extensively used to demonstrate model performanceand are easily accessible. Our attack is simple and straightforward. Weleverage the shadow model training techniques to generate training data for theattack model which is essentially an image classifier. Extensive evaluation onthree benchmark datasets shows that our proposed attack can effectively inferthe architecture/hyperparameters of image classifiers based on convolutionalneural network (CNN) given the scientific plot generated from it. We alsoreveal that the attack's success is mainly caused by the shape of thescientific plots, and further demonstrate that the attacks are robust invarious scenarios. Given the simplicity and effectiveness of the attack method,our study indicates scientific plots indeed constitute a valid side channel formodel information stealing attacks. To mitigate the attacks, we propose severaldefense mechanisms that can reduce the original attacks' accuracy whilemaintaining the plot utility. However, such defenses can still be bypassed byadaptive attacks.

A Plot is Worth a Thousand Words: Model Information Stealing Attacks via Scientific Plots