Project: Re-Use Code from Literature

We are studying adversarial examples and wanted to use the code presented in the orignal paper and also in a PyTorch example/ a ML Notebook.
Our project focuses on understanding how this literature attack affects three different optimisers, before and after training neural nets with adversial examples.
We wondering if we have to recreate the attack code or whether we can reuse the code that already existed and just cite the source?
We figured that our focus was on how it affects our optimisers and not on creating an attack itself this would be okay?

Top comment

sure, it's highly recommended to reuse (and cite) existing code. such as
https://github.com/RobustBench/robustbench
or
https://github.com/tml-epfl/adv-training-corruptions
or others...

Page 1 of 1

Add comment

Post as Anonymous Dont send out notification