Tutorial 10: Adversarial Attacks (Part 1) скачать в хорошем качестве

Tutorial 10: Adversarial Attacks (Part 1)

In this tutorial, we will discuss adversarial attacks on deep image classification models. As we have seen in many of the previous tutorials so far, Deep Neural Networks are a very powerful tool to recognize patterns in data, and, for example, perform image classification on a human-level. However, we have not tested yet how robust these models actually are. Can we “trick” the model and find failure modes? Can we design images that the networks naturally classify incorrectly? Due to the high classification accuracy on unseen test data, we would expect that this can be difficult. However, in 2014, a research group at Google and NYU showed that deep CNNs can be easily fooled, just by adding some salient but carefully constructed noise to the images. We will implement simple white-box attacks ourselves, including the Fast Gradient Sign Method (FGSM) and Adversarial Patches. This notebook is part of a lecture series on Deep Learning at the University of Amsterdam. The full list of tutorials can be found at https://uvadlc-notebooks.rtfd.io. Link to the notebook: https://uvadlc-notebooks.readthedocs.... 00:00 Intro 03:14 Deep CNNs on ImageNet 06:15 White-box adversarial attacks 07:44 Fast Gradient Sign Method