Machine learning is a powerful tool with many well-suited applications for malware detection, classification, and risk quantification. Despite its reputation as a "black box" component to an enterprise security solution, designing a robust machine learning model for malware detection is an involved process: its success hinges on understanding the problem you're trying to solve, the underlying data you utilize, and most importantly, its limitations.
In this session, we analyze working models discuss the strengths, pitfalls, and high-level trade-offs of using machine learning for successful malware detection.
Frank Jas is currently Cyphort's Chief Architect. He was the number three employee at Cyphort and has been responsible for designing and building many components of Cyphort's products, including the detonation engines, machine learning-based analytics, malicious PDF detector, clustering, the UI and HTTP API.