I just completed this course in Berkeley and I can’t thank the instructors enough for such a wonderful course and website. Does anyone know of a good resource where I can read more about deep learning in adversarial setting, ways to attack neural networks and possible defenses? Even better, I hope the authors of this book can consider the addition of this topic
Great to hear that you enjoyed the class.
This class at MIT https://people.csail.mit.edu/madry/6.883/ covers quite a bit about adversarial attacks in deep learning.
There’s this paper by GoodFellow https://arxiv.org/abs/1412.6572 and this survey paper as well https://arxiv.org/pdf/1810.00069.pdf about adversarial attacks. And this one about a way to resist them: https://openreview.net/pdf?id=S18Su--CW