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

Converting malware into images followed by vision-based deep learningalgorithms has shown superior threat detection efficacy compared with classicalmachine learning algorithms. When malware are visualized as images,visual-based interpretation schemes can also be applied to extract insights ofwhy individual samples are classified as malicious. In this work, via two casestudies of dynamic malware classification, we extend the local interpretablemodel-agnostic explanation algorithm to explain image-based dynamic malwareclassification and examine its interpretation fidelity. For both case studies,we first train deep learning models via transfer learning on malware images,demonstrate high classification effectiveness, apply an explanation method onthe images, and correlate the results back to the samples to validate whetherthe algorithmic insights are consistent with security domain expertise. In ourfirst case study, the interpretation framework identifies indirect calls thatuniquely characterize the underlying exploit behavior of a malware family. Inour second case study, the interpretation framework extracts insightfulinformation such as cryptography-related APIs when applied on images createdfrom API existence, but generate ambiguous interpretation on images createdfrom API sequences and frequencies. Our findings indicate that currentimage-based interpretation techniques are promising for explaining vision-basedmalware classification. We continue to develop image-based interpretationschemes specifically for security applications.

To believe or not to believe: Validating explanation fidelity for dynamic malware analysis