YiZeng623/frequency-backdoor

ICCV 2021, We find most existing triggers of backdoor attacks in deep learning contain severe artifacts in the frequency domain. This Repo. explores how we can use these artifacts to develop stronger backdoor defenses and attacks.

36
/ 100
Emerging

This project helps deep learning security researchers and practitioners identify and understand 'backdoor' vulnerabilities in image recognition models. It takes trained models or image datasets and reveals hidden patterns, called 'triggers,' that attackers might use. The output helps in developing ways to detect these attacks or create more stealthy ones.

No commits in the last 6 months.

Use this if you are a deep learning security researcher or practitioner working to protect or test image recognition models against hidden, malicious triggers.

Not ideal if you are looking for a general-purpose image analysis tool unrelated to deep learning security or backdoor attacks.

deep-learning-security backdoor-attack-detection neural-network-robustness computer-vision-security adversarial-machine-learning
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 16 / 25
Community 12 / 25

How are scores calculated?

Stars

48

Forks

6

Language

Jupyter Notebook

License

MIT

Last pushed

Apr 27, 2022

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/YiZeng623/frequency-backdoor"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.