PulseAugur
EN
LIVE 05:58:21

New DROIDBREAKER framework creates functional adversarial Android malware

By PulseAugur Editorial · [2 sources] ·

Researchers have developed DROIDBREAKER, a new framework designed to create practical and functional adversarial Android applications (APKs) that can evade machine-learning malware detectors. This framework addresses limitations in existing methods, which are often impractical due to build failures or semantic unreliability. DROIDBREAKER employs query-efficient attacks by manipulating influential APK components and uses fine-grained, build-safe modifications to preserve the application's core functionality, as validated by runtime equivalence testing. AI

IMPACT This research highlights vulnerabilities in ML-based malware detection, potentially necessitating more robust security measures in Android application development.

RANK_REASON The cluster contains a research paper detailing a new framework for adversarial attacks on machine learning models.

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New DROIDBREAKER framework creates functional adversarial Android malware

COVERAGE [2]

  1. arXiv cs.LG TIER_1 English(EN) · Christian Scano, Diego Soi, Angelo Sotgiu, Luca Demetrio, Davide Maiorca, Giorgio Giacinto, Fabio Roli, Battista Biggio ·

    DroidBreaker: Practical and Functional Problem-Space Attacks on Machine-Learning Android Malware Detectors

    arXiv:2606.26707v1 Announce Type: cross Abstract: Adversarial APKs are Android applications modified in the problem space to evade machine-learning malware detectors. In this work, we first show that, despite claims, existing problem-space attacks remain largely impractical. Most…

  2. arXiv cs.LG TIER_1 English(EN) · Battista Biggio ·

    DroidBreaker: Practical and Functional Problem-Space Attacks on Machine-Learning Android Malware Detectors

    Adversarial APKs are Android applications modified in the problem space to evade machine-learning malware detectors. In this work, we first show that, despite claims, existing problem-space attacks remain largely impractical. Most techniques leverage software transplantation to i…

RELATED ENTITIES

RELATED TOPICS