Home

Awesome

Mitigating Adversarial Effects Through Randomization

This paper proposed to utilize randomization to mitigate adversarial effects (https://arxiv.org/pdf/1711.01991.pdf). By combining the proposed randomization method with an adversarially trained model, it ranked No.2 among 107 defense teams in the NIPS 2017 adversarial examples defense challenge (https://www.kaggle.com/c/nips-2017-defense-against-adversarial-attack).

The approach

The main ideal of the defense is to utilize randomization to defend adversarial examples:

In general, the pipeline is shown below:

Pipeline

Pros

  1. No additional training/finetuning is required
  2. Very little computation introduced
  3. Compatiable to different networks and different defending methods (i.e., we use randomization + ensemble adversarial training + Inception-Resnet-v2 in our submission)

Ensemble adversarial training model

Team Member

Leaderboard

Our team name is iyswim, and our rank is No.2.

Citing this work

If you find this work is useful in your research, please consider citing:

@inproceedings{xie2017mitigating,
    title={Mitigating Adversarial Effects Through Randomization},
    author={Xie, Cihang and Wang, Jianyu and Zhang, Zhishuai and Ren, Zhou and Yuille, Alan},
    booktitle={International Conference on Learning Representations},
    year={2018}
}