Loops is a tool for generating lots of untargeted adversarial examples, created as part of my bachelor's thesis.

This is proof of concept code. It's not meant to be integrated with other tools.

Lots of techiniques for adversarial examples are aimed are minimizing distance metrics such as L2 and L∞. This is good in theory, because it allows for an easy comparison between different attacks. In practice however, there are scenarious where these measures are irrelevant.

This version uses simple chaining with a subset of attack algorithms available in Foolbox.

Some code examples can be found in the project wiki.