Skip to content

LucasSloan/SphereDiscriminator

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 

Repository files navigation

SphereDiscriminator

My recreation of Ian Goodfellow's Adversarial Spheres paper.

sphere_discriminator.py trains a model to predict if a point is on a sphere of radius 1 or a sphere or radius 1.3, then uses gradient descent to try to find a point on a sphere of radius 1 that the model misclassifies. In accordance with the paper's findings, in lower dimensions, it is extremely difficult to produce adversarial examples. Conversely, in higher dimensions, adversarial examples are easy to produce.

Run with:

python sphere_discriminator.py <Number of Dimensions>

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages