Adversarial Robustness

This video is part of the Introduction to ML Safety course (https://course.mlsafety.org) and was recorded by Dan Hendrycks at the Center for AI Safety (https://safe.ai). Check out our other projects: Community: https://mlsafety.org Newsletter: https://newsletter.mlsafety.org/ Twitter:   / ml_safety   Reddit:   / mlsafety   This video covers the following topics in adversarial robustness: Optimization pressure Projected gradient attack (PGD) Untargeted vs targeted attacks Adversarial evaluation White box vs black box attacks Transferability Unforeseen attacks Text attacks Robustness certificates