Explore the concept of Deepfool, a method for generating adversarial examples in deep neural networks, in this 20-minute lecture from the University of Central Florida. Learn about linearizing classifiers, multiclass classification, and experimental results. Examine the objectives, implementation, and potential drawbacks of this technique in the context of computer vision and machine learning security.
Computer Vision - Adversarial Attacks and Deepfool Algorithm - Lecture 1
University of Central Florida via YouTube
Overview
Syllabus
Introduction
Summary
Objectives
Deepfool
Linearizing a classifier
Multiclass classification
Experiments
Conclusion
Disadvantages
Taught by
UCF CRCV
