PEsidious - Create Mutated Evasive Malware Using AI

Explore the cutting-edge research on using Artificial Intelligence to create mutated evasive malware samples in this conference talk from the Hack In The Box Security Conference. Dive into the innovative PEsidious project, which combines Generative Adversarial Networks and Deep Reinforcement Learning to generate adversarial feature vectors and teach models to perform additional mutations that reduce malware detection rates. Learn about the various mutation techniques applied to imports, exports, headers, signatures, sections, and size. Discover how this approach overcomes limitations of previous research and brings together different AI models for improved results. Understand the potential applications of this technology for both offensive and defensive cybersecurity strategies. Gain insights from the creators of PEsidious, who won first place and $40,000 in the HITB CyberWeek AI challenge 2019, as they present additional features to enhance its efficiency and impact on the cybersecurity landscape.