jiangw-0/LE_JCDP
Unlearnable Examples Give a False Sense of Security: Piercing through Unexploitable Data with Learnable Examples
This project helps machine learning researchers evaluate the robustness of 'unlearnable examples' – data designed to prevent models from learning specific information. It takes existing poisoned image datasets (unlearnable examples) and pre-trained diffusion models as input. The output is 'learnable examples' that can be used to test and bypass the intended unlearnability, revealing vulnerabilities in data poisoning defenses. This tool is for researchers focusing on adversarial machine learning and data privacy.
No commits in the last 6 months.
Use this if you are a machine learning researcher who wants to test the effectiveness of unlearnable examples and understand how to make them learnable again.
Not ideal if you are looking for a plug-and-play solution for general image generation or if you are not deeply involved in adversarial machine learning research.
Stars
11
Forks
—
Language
Python
License
—
Category
Last pushed
Oct 14, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/jiangw-0/LE_JCDP"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
OPTML-Group/Unlearn-Saliency
[ICLR24 (Spotlight)] "SalUn: Empowering Machine Unlearning via Gradient-based Weight Saliency in...
Shilin-LU/VINE
[ICLR 2025] "Robust Watermarking Using Generative Priors Against Image Editing: From...
WindVChen/DiffAttack
An unrestricted attack based on diffusion models that can achieve both good transferability and...
koninik/DiffusionPen
Official PyTorch Implementation of "DiffusionPen: Towards Controlling the Style of Handwritten...
Wuyxin/DISC
(ICML 2023) Discover and Cure: Concept-aware Mitigation of Spurious Correlation