ByungKwanLee/Causal-Adversarial-Instruments
[CVPR 2023] Official PyTorch Implementation for "Demystifying Causal Features on Adversarial Examples and Causal Inoculation for Robust Network by Adversarial Instrumental Variable Regression"
This project helps machine learning engineers and researchers build more robust image classification models that are resistant to 'adversarial attacks.' It takes an existing image dataset and a deep neural network, then applies specialized training techniques to identify and strengthen the model's 'causal features' — the true reasons it makes a correct prediction — making it less susceptible to small, intentionally crafted input changes that can fool other models. The output is a more reliable and secure image classification model.
No commits in the last 6 months.
Use this if you need to develop deep learning models for image classification that can withstand malicious tampering or 'adversarial examples' in real-world applications.
Not ideal if your primary goal is general image classification performance without specific concerns about adversarial robustness or if you are not familiar with deep learning model training.
Stars
45
Forks
5
Language
Python
License
MIT
Category
Last pushed
Jul 18, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/ByungKwanLee/Causal-Adversarial-Instruments"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.