Completed
What Are the Adversary's Capabilities? To generate attacks, attacker needs to know how changing input affects output
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Accessorize to a Crime - Real and Stealthy Attacks on State-Of-The-Art Face Recognition
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Machine Learning Is Ubiquitous
- 3 What Do You See?
- 4 The Difference
- 5 What Are the Adversary's Capabilities? To generate attacks, attacker needs to know how changing input affects output
- 6 What's a (Deep) Neural Network?
- 7 Face Recognition . Applications: surveillance, access control...
- 8 Face Recognition: Our Attacks
- 9 Deep Face Recognition
- 10 Apply Changes to Face Only
- 11 Apply Changes to Eyeglasses
- 12 Experiments in Digital Environment
- 13 Smooth Transitions Natural images tend to be smooth
- 14 Printable Eyeglasses Chalenge: Cannot print all colors
- 15 Robust Perturbations
- 16 Putting All the Pieces Together - Physically realizable impersonation
- 17 Does This Work?
- 18 Experiment: Realized Impersonations
- 19 Impersonation Attacks Pose Real Risk!
- 20 Extensions (See Paper)
- 21 Conclusions
