AppWalker: Efficient and Accurate Dynamic Analysis of Apps via Concolic Walking Along the Event-Dependency Graph

摘要

Dynamic analyzing techniques play an important and unique role in detecting Android malware and vulnerabilities, as they can provide higher precision than static methods. However, they are inherently incomplete and inefficiency. We attack this problem by proposing a novel method, i.e., concolic walking along the event-dependency graph. We implement AppWalker based on it. Evaluation over a real-life app set shows that better efficiency and accuracy than state-of-the-art concolic analysis tools are achieved.