Overview
Explore the vulnerabilities in machine learning systems for speech, text, and face recognition in this 50-minute conference talk from BSidesLV 2018. Delve into the methods used to hack ML algorithms across various applications as presenters Guy Barnhart-Magen and Ezra Caltum reveal how even advanced AI systems like JARVIS can be compromised. Gain insights into the potential security risks associated with ML technologies and learn about the importance of robust defenses in an increasingly AI-driven world.
Syllabus
BG - JARVIS Never Saw It Coming: Hacking machine learning (ML) in Speech, Text and Face Recognition
Taught by
BSidesLV