Implementation of Adversarial Attacking Tools

The project is a part of UW ECE653, Software Testing, Quality Assurance and Maintenance. It involves:

• Attacking state-of-the-art imagenet models with FDSM, PGD, and DeepFool methods
• Using modern attacking tools to generate attacks on these models
• Extending a tool with new attacking method

Attacks Performed

• Fast Gradient Sign Method
• Projected Gradient Descent
• DeepFool

Tool Used

• CleverHans
• ART
• FoolBox
• AdvBox
• ARES

New Attack Methods Implemented

• DeepFool Attack for CleverHans. Can be found here