Explore a novel approach to malware capability detection in this Black Hat USA 2013 conference talk. Learn about CrowdSource, an open-source machine learning-based reverse engineering tool that leverages millions of technical documents from the web to identify high-level malware functionality. Discover how this DARPA Cyber Fast Track-funded project aims to provide rapid, automated analysis of malware capabilities, including the ability to detect features like screenshot capture, IRC communication, and webcam operation. Gain insights into the tool's innovative features, such as probabilistic capability detection and traceable output with web document citations. Examine the algorithm behind CrowdSource, its training process using web data, and compelling results demonstrating its effectiveness in reverse engineering active malware variants. Understand the potential impact of this tool on improving visibility into the global malware landscape and accelerating the malware analysis process for security practitioners.
Read more
CrowdSource - Crowd Trained Machine Learning Model for Malware Capability Detection