Tutorial 10 Adversarial Attacks Part 1

Summary

15780 Final Project Sergio Claure Xiaohui Du. This is the video illustration of the year-long project - " Defense mechanism against patch Presenters: Han Xu, Yaxin Li, Wei Jin, Jiliang Tang (Michigan State University) We first discuss the mismatch between the train and inference time distributions. We explain that an adversary might present to the ... Find out how to fool a neural network. 00:00 Introduction 02:29 Classification Loss 08:19 Project Webpage: light.princeton.edu/ Existing neural networks for computer vision tasks are vulnerable to