Explore the world of adversarial machine learning in this comprehensive conference talk from Conf42 ML 2024. Delve into various types of attacks, including poisoning, property inference, membership inference, model extraction, and evasion. Examine real-world examples such as the Tay chatbot incident, PoisonGPT, and attacks on Tesla's autopilot and object detection systems. Learn about the OWASP Top 10 for large language models and discover effective mitigation strategies to protect against adversarial machine learning threats. Gain valuable insights into this critical aspect of AI security and its implications for the future of machine learning.
Overview
Syllabus
intro
preamble
about me
machine learning introduction
what is adversarial machine learning?
types of adversarial machine learning attacks
poisoning attack
tay chatbot
poisongpt
property inference attack
membership inference attack
model extraction attack
evasion attack
invisibiliyu cloak
attacking tesla's autopilot
attacking object detection systems
owasp top 10 for large language models
mitigation strategies
mitigating adversarial ml attacks
summary
my contact information
Taught by
Conf42
Related Courses
-
Zen and the Art of Adversarial Machine Learning
-
Membership Inference Attacks against Machine Learning Models
-
Adversary Instantiation - Lower Bounds for Differentially Private Machine Learning
-
Adversarial and Poisoning Attacks Against Speech Systems - Where to Find Them
-
Trusted and Responsible AI - Explainability, Adversarial Attacks, Bias, and Fairness
-
Adversarial Examples in Machine Learning - Crafting and Defending Against Attacks
