BACKDOORBOX: A PYTHON TOOLBOX FOR BACKDOOR LEARNIN
Contribution analysis of BackdoorBox.
Main contribution
GOAL
: To facilitate the research and development of more secure training schemes and defenses.
There are four main characteristic of the BackdoorBox.
Toolbox characteristics
- Consistency: reimplement all methods in a unified manner.
- Simplicity: provide code example explain how to use them, and with necessary code comments.
- Flexibility: gain main components easily, such as poisoned dataset, implemented attaks and defenses.
- Co-development: open-source
Backdoor attack defination
Categorize existing backdoor attacks into three main types:
- poinson-only backdoor attacks
- trining-controlled backdoor attacks
- model-modified backdoor attacks
Backdoor defense defination
Categorize existing backdoor defenses into six main types:
- pre-processin-based defenses
- model repairing
- poison suppression
- model diagnosis
- sample diagnosis
- certified defenses
Conclusion
Mainly developed for flexibale use.
All articles in this blog are licensed under CC BY-NC-SA 4.0 unless stating additionally.