Machine Learning Security University Project

Re-evaluate 5 RobustBench models with another attack algorithm (e.g. FMN) and identify samples for which one attack works and the other doesn't. Explain the results - i.e., provide some motivations on why one of the attacks did not work properly, while the other did.