This repo serves as the reimplementation and extension for the paper The Curious Case of Adversarially Robust Models: More Data Can Help, Double Descend, or Hurt Generalization.
Reports of all our experiments can be found in the Contribution folder.
“Reproducability Report” reproduces all the results of the original paper.
"Extension_notebook" summarizes the results of all our extension experiments with multi-dimensional data and various different attacks.
"Extension_report" is the sum of all our extension ideas.
All remaining python files are experimental models.