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

In recent years, the topic of explainable machine learning (ML) has beenextensively researched. Up until now, this research focused on regular ML usersuse-cases such as debugging a ML model. This paper takes a different postureand show that adversaries can leverage explainable ML to bypass multi-featuretypes malware classifiers. Previous adversarial attacks against suchclassifiers only add new features and not modify existing ones to avoid harmingthe modified malware executable's functionality. Current attacks use a singlealgorithm that both selects which features to modify and modifies them blindly,treating all features the same. In this paper, we present a different approach.We split the adversarial example generation task into two parts: First we findthe importance of all features for a specific sample using explainabilityalgorithms, and then we conduct a feature-specific modification,feature-by-feature. In order to apply our attack in black-box scenarios, weintroduce the concept of transferability of explainability, that is, applyingexplainability algorithms to different classifiers using different featuressubsets and trained on different datasets still result in a similar subset ofimportant features. We conclude that explainability algorithms can be leveragedby adversaries and thus the advocates of training more interpretableclassifiers should consider the trade-off of higher vulnerability of thoseclassifiers to adversarial attacks.

Generating End-to-End Adversarial Examples for Malware Classifiers Using Explainability