Explore a 22-minute IEEE conference talk that delves into the concept of BadEncoder, focusing on backdoor attacks to pre-trained encoders in self-supervised learning. Learn from Duke University researchers Jinyuan Jia, Yupei Liu, and Neil Zhenqiang Gong as they present their findings on this critical cybersecurity topic. Gain insights into the vulnerabilities of pre-trained encoders and understand the potential risks associated with self-supervised learning techniques. Discover the implications of these backdoor attacks and their impact on the field of machine learning and artificial intelligence.
BadEncoder - Backdoor Attacks to Pre-trained Encoders in Self-Supervised Learning
Overview
Syllabus
BadEncoder: Backdoor Attacks to Pre-trained Encoders in Self-Supervised Learning
Taught by
IEEE Symposium on Security and Privacy
Tags
Related Courses
-
Secure Self-supervised Learning: Challenges and Solutions
-
RAB - Provable Robustness Against Backdoor Attacks
-
GeeSolver - A Generic, Efficient, and Effortless Solver with Self-Supervised Learning for Breaking Text Captchas
-
Selective Amnesia - On Efficient, High-Fidelity and Blind Suppression of Backdoor Effects in Trojaned Machine Learning Models
-
Backdoor Attacks and Defenses in AI Systems - Seminar Series
-
REDEEM MYSELF - Purifying Backdoors in Deep Learning Models Using Self Attention Distillation
