This repo is aimed to build a library containing codes of popular backdoor attacks and defenese. The main goal is to make the implementation simple and easy-to-use, meanwhile without learning how the code is writing.
Current codes containing:
- Defense: SentiNet: Detecting Localized Universal Attacks Against Deep Learning Systems. I cannot found an official implementation.
- More attacks and defense: coming...
- 2023/3/22: Add defense, SentiNet.
- SentiNet: Please refer to the blog: https://weichengan.com/2023/03/22/suibi/sentinet_pytorch/.