• K-Arm Backdoor Optimization
• • Setup Environments
  • • Install Anaconda Python
    • Setup the Conda Environment
  • Datasets
  • • TrojAI
    • ImageNet
  • Quick Start
  • Results
  • Contacts

K-Arm Backdoor Optimization

This is the official repository of the ICML2021 paper Backdoor Scanning for Deep Neural Networks through K-Arm Optimization in PyTorch.

If you find this code is useful for your research, please cite the following:

@article{shen2021backdoor,
  title={Backdoor Scanning for Deep Neural Networks through K-Arm Optimization},
  author={Shen, Guangyu and Liu, Yingqi and Tao, Guanhong and An, Shengwei and Xu, Qiuling and Cheng, Siyuan and Ma, Shiqing and Zhang, Xiangyu},
  journal={arXiv preprint arXiv:2102.05123},
  year={2021}
}

Setup Environments

We suggest to use Conda for testing the code on TrojAI datasets. Detailed instruction can be found here.

Install Anaconda Python

https://www.anaconda.com/distribution/

Setup the Conda Environment

1. conda create --name trojai-example python=3.8 -y

2. conda activate trojai-example

3. Install required packages into this conda environment

   1. conda install pytorch=1.7.0 torchvision=0.8.0 torchtext==0.8.0 cudatoolkit=11.0 -c pytorch -c conda-forge
   2. pip install --upgrade trojai
   3. conda install jsonpickle

Datasets

TrojAI

The code is tested on TrojAI datasets (round1-4). TrojAI datasets can be accessed at TrojAI website

ImageNet

We will release the ImageNet pre-trained models and code in the near future.

Quick Start

To test the code on TrojAI datasets, simply run command

$ python main.py --result_filepath <resultFilepath> --examples_dirpath <dataDirpath> --model_filepath <modelFilepath>

To run the code on custom models, make sure your sample images in the <examples_dirpath> have the following format:

examples_dirpath/class_<class_id>_example_<example_id>.png

Description about the main parameters:

• <Beta>: Coefficient in the K-Arm schedulor objective function
• <gamma,global(local)_theta>: Parameters in the Arm Pre-screening procedure
• <global(local,ratio)_det_bound>: Trigger size bound for detecting different types of backdoors
• <epsilon_for_bandits>: Controls the randomness during the K-Arm optimization

Results

Please check our results here under the team name Perspecta-PurdueRutgers. The default settings of the parameters in this repo can achieve 90% detection accuracy on TrojAI round3 training and testing datasets. For scanning different types of models, some parameters might need tunning.

Contacts

Guangyu Shen, shen447@purdue.edu
Yingqi Liu, liu1751@purdue.edu