Long-term Cross Adversarial Training

Structure

1. Few-shot-datasets folder: store datasets including MiniImageNet, TieredImageNet, CIFAR-FS
2. data folder: codes to preprocess and load dataset
3. models folder and qpth model: embedding network
4. experiments folde: store checkpoints, test results, log file, figures
5. test.py: test model
6. train.py: train model
7. utils.py: utilities

Requirements

• Pytorch
• Python
• CUDA
• Numpy
• Matplotlib

Datasets and models

• https://github.com/kjunelee/MetaOptNet
• https://github.com/goldblum/AdversarialQuerying

Credits

• Goldblum, M., Fowl, L., and Goldstein, T. Adversarially robust few-shot learning: A meta-learning approach.Advances in Neural Information Processing Systems, 33,2020.
• Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun Deep Residual Learning for Image Recognition. arXiv:1512.03385
• Cihang Xie, Yuxin Wu, Laurens van der Maaten, Alan Yuille, Kaiming He Feature Denoising for Improving Adversarial Robustness. arXiv:1812.03411

License

• MIT License

Cite ours

@inproceedings{
liu2021longterm,
title={Long-term Cross Adversarial Training: A Robust Meta-learning Method for Few-shot Classification Tasks },
author={Fan Liu and Shuyu Zhao and Xuelong Dai and Bin Xiao},
booktitle={ICML 2021 Workshop on Adversarial Machine Learning},
year={2021},
url={https://openreview.net/forum?id=RVlevnrbjnU}
}