Overview
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Explore a thought-provoking keynote address delivered at the 1st Deep Learning and Security Workshop during the 2018 IEEE Symposium on Security & Privacy. Delve into the intriguing question "Is 'Adversarial Examples' an Adversarial Example?" presented by David Evans. Examine topics such as evolutionary search, genetic programming, defense strategies, human perception, and the evaluation of defenses against adversarial attacks. Gain insights into the computational costs, greedy strategies, and the differences between blackbox and whitebox approaches. Conclude with reflections on the implications of adversarial examples in deep learning and security.
Syllabus
Introduction
It is a momentous day
DDP are the reason
Law Article 22
Evolutionary Search
Genetic Programming
Defense
Human Perception
Minimize
Random Noise
Feature Squeeze
How should we evaluate defenses
The computational cost
Simple greedy strategy
Blackbox vs Whitebox
Madre Defense
Conclusion
Ali Rahimi
Young Lukens
Taught by
IEEE Symposium on Security and Privacy