Targeted Edge-informed Attack (TEA)

This repository is associated with the paper titled "Accelerating Targeted Hard-Label Adversarial Attacks in Low-Query Black-Box Settings".

Contents

• ATTACK/: Core implementation and a detailed Jupyter notebook tutorial (ATTACK/tutorial.ipynb).
• ILSVRC2012_val_pairs/: Five example image pairs from the ImageNet validation dataset used for the notebook tutorial.

Citation

Please consider citing our work if it is beneficial to your research.

License

This project is released under the MIT License. See LICENSE for details.

Contact for Questions