Introduction to adversarial examples - Deepl learning
- Changing the prediction result of an image classification without changing its visual to the human eye
- Python 3.7+
- jupyter ~= 1.0.0
- matplotlib ~= 3.3
- tensorflow ~= 2.6
How to install all the requirements :
pip3 install -r requirements.txt
jupyter notebook Adversarial\ examples\ -\ Deep\ learning.ipynb
Changing the result of the prediction to a random class (Non-directed), class 3 to 5 (random)
Changing the result of the prediction to a specific class (Directed), class 3 to 9
.
├── Adversarial examples - Deep learning.ipynb
├── imgs
│ ├── result1.png
│ └── result.png
├── README.md
└── requirements.txt
- Luis Rosario - Member 1 - Luisrosario2604