PhD dissertation presenting scalable semantics-based approaches for detecting similar Android applications, with applications to clone detection, malware analysis, and security assessment.
Journal article presenting AnDarwin’s scalable semantic analysis approach that detected 4,295 cloned and 36,106 rebranded apps from 265,359 applications across 17 markets including Google Play, with automatic library code removal.
Presents MAdFraud, analyzing 130,339 apps to detect mobile ad fraud including background ad requests (30% of apps) and 27 apps with automated clicking, revealing stealthy techniques to evade detection.
Introduces AnDarwin, a scalable tool that analyzed 265,359 apps from 17 markets to detect 4,295 clones and 36,106 rebranded apps, discovering 88 new malware variants by comparing semantic information without pairwise comparison.
First large-scale study characterizing plagiarized Android apps across 265,359 applications from 17 markets and estimating their economic impact by analyzing advertising revenue siphoned from original developers using tier-1 cellular network traces.
Master’s thesis investigating Android application plagiarism through program analysis techniques to detect cloned and repackaged mobile apps.
Presents DNADroid, a tool that detects Android application cloning by comparing program dependency graphs, identifying at least 141 cloned apps including cases of malware injection and ad revenue redirection.
Presents AndroidLeaks, which analyzed 24,350 Android apps in 30 hours and found 2,342 apps leaking private data including phone info, GPS location, WiFi data, and audio from 57,299 potential privacy leaks detected.
Examines privacy implications of 13 Android ad libraries, discovering several checked dangerous permissions beyond documentation (CAMERA, WRITE_CALENDAR, WRITE_CONTACTS) and identifying cross-app user tracking capabilities.