HanxunH/CognitiveDistillation
[ICLR2023] Distilling Cognitive Backdoor Patterns within an Image
This project helps identify 'backdoor patterns' hidden within images that could manipulate a pre-trained AI model. By inputting a trained image classification model and a batch of images, it outputs 'masks' that highlight these suspicious regions. This tool is for AI security researchers or model auditors concerned with detecting and analyzing poisoned data.
Use this if you need to detect hidden, malicious patterns in images that could cause your AI models to behave unexpectedly or incorrectly.
Not ideal if you are looking for a general image anomaly detection tool or a method to improve model accuracy.
Stars
36
Forks
3
Language
Python
License
MIT
Category
Last pushed
Oct 29, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/HanxunH/CognitiveDistillation"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
QData/TextAttack
TextAttack 🐙 is a Python framework for adversarial attacks, data augmentation, and model...
ebagdasa/backdoors101
Backdoors Framework for Deep Learning and Federated Learning. A light-weight tool to conduct...
THUYimingLi/backdoor-learning-resources
A list of backdoor learning resources
zhangzp9970/MIA
Unofficial pytorch implementation of paper: Model Inversion Attacks that Exploit Confidence...
LukasStruppek/Plug-and-Play-Attacks
[ICML 2022 / ICLR 2024] Source code for our papers "Plug & Play Attacks: Towards Robust and...