Skip to content

ain-soph/trojanzoo

Repository files navigation

TrojanZoo

logo

contact License

python>=3.11 docs

release pypi docker

NOTE: TrojanZoo requires python>=3.11, pytorch>=2.0.0 and torchvision>=0.15.0, which must be installed manually. Recommend to use conda to install.

This is the code implementation (pytorch) for our paper in EuroS&P 2022:
TrojanZoo: Towards Unified, Holistic, and Practical Evaluation of Neural Backdoors

TrojanZoo provides a universal pytorch platform to conduct security researches (especially backdoor attacks/defenses) of image classification in deep learning. It is composed of two packages: trojanzoo and trojanvision. trojanzoo contains abstract classes and utilities, while trojanvision contains abstract and concrete ones for image classification task.

Note: This repository is also maintained to cover the implementation of
our kdd 2020 paper AdvMind: Inferring Adversary Intent of Black-Box Attacks
and ccs 2020 paper A Tale of Evil Twins: Adversarial Inputs versus Poisoned Models

Documentation

We have documentation available at https://ain-soph.github.io/trojanzoo.

Screenshot

screenshot

Features

  1. Colorful and verbose output!

    Note: enable with --color for color and --verbose for verbose.
    To open an interactive window with color, use python - --color

  2. Modular design (plug and play)
  3. Good code linting support
  4. Register your own module to the library.
  5. Native Pytorch Output
    trojanzoo and trojanvision provides API to generate raw pytorch instances, which makes it flexible to work with native pytorch and other 3rd party libraries.

    trojanzoo.datasets.DataSet can generate torch.utils.data.Dataset and torch.utils.data.DataLoader
    trojanzoo.models.Model attribute _model is torch.nn.Module, attribute model is torch.nn.DataParallel
    Specifically, trojanvision.datasets.ImageSet can generate torchvision.datasets.VisionDataset, trojanvision.datasets.ImageFolder can generate torchvision.datasets.ImageFolder

  6. Enable pytorch native AMP(Automatic Mixed Precision) with --amp for training
  7. Flexible Configuration Files
  8. Good help information to check arguments. (-h or --help)
  9. Detailed and well-organized summary() for each module.

Installation

  1. pip install trojanzoo
  2. pip install --upgrade git+https://github.com/ain-soph/trojanzoo.git
  3. (HIGHLY RECOMMEND)
    git clone https://github.com/ain-soph/trojanzoo
    pip install -e trojanzoo
    

    This could install the github repo as a package but avoid copying files to site_packages, so that you can easily keep it updated by doing git pull.

  4. docker pull local0state/trojanzoo or docker pull ghcr.io/ain-soph/trojanzoo

Quick Start

You can use the provided example scripts to reproduce the evaluation results in our paper.

Note: The program won't save results without --save

  1. Train a model:
    e.g. ResNet18 on CIFAR10 with 95% Acc

    python ./examples/train.py --color --verbose 1 --dataset cifar10 --model resnet18_comp --lr_scheduler --cutout --grad_clip 5.0 --save
  2. Test backdoor attack (e.g., BadNet):
    e.g. BadNet with ResNet18 on CIFAR10

    python ./examples/backdoor_attack.py --color --verbose 1 --pretrained --validate_interval 1 --dataset cifar10 --model resnet18_comp --attack badnet --mark_random_init --epochs 50 --lr 0.01 --save
  3. Test backdoor defense (e.g., Neural Cleanse):
    e.g. Neural Cleanse against BadNet

    python ./examples/backdoor_defense.py --color --verbose 1 --pretrained --validate_interval 1 --dataset cifar10 --model resnet18_comp --attack badnet --defense neural_cleanse --mark_random_init --epochs 50 --lr 0.01

IMC

python ./examples/backdoor_attack.py --color --verbose 1 --pretrained --validate_interval 1 --dataset cifar10 --model resnet18_comp --attack imc --mark_random_init --epochs 50 --lr 0.01 --save

AdvMind

(with attack adaptive and model adaptive)

python ./examples/adv_defense.py --color --verbose 1 --pretrained --validate_interval 1 --dataset cifar10 --model resnet18_comp --attack pgd --defense advmind --attack_adapt --defense_adapt

Detailed Usage

Configuration file structure

All arguments in the parser are able to set default values in configuration files.
If argument values are not set in the config files, we will use the default values of __init__()

Parameters Config: (priority ascend order)

The higher priority config will override lower priority ones.
Within each priority channel, trojanvision configs will overwrite trojanzoo

  1. Package Default: /trojanzoo/configs/, /trojanvision/configs/

    These are package default settings. Please don't modify them.
    You can use this as a template to set other configs.

  2. User Default: ~/.trojanzoo/configs/trojanzoo/, ~/.trojanzoo/configs/trojanvision/
  3. Workspace Default: /configs/trojanzoo/, /configs/trojanvision/
  4. Custom Config: --config [config location]
  5. CMD parameters: --[parameter] [value]

Store path of Dataset, Model, Attack & Defense Results

Modify them in corresponding config files and command-line arguments.

Dataset: --data_dir (./data/data)
Model: --model_dir (./data/model)
Attack: --attack_dir (./data/attack)
Defense: --defense_dir (./data/defense)

Output Verbose Information:

  1. CMD modules: --verbose 1
  2. Colorful output: --color
  3. tqdm: --tqdm
  4. Check command-line argument usage: --help
  5. AdvMind verbose information: --output [number]

Use your DIY Dataset/Model/Attack/Defense

  1. Follow our example to write your DIY class. (CIFAR10, ResNet, IMC, Neural Cleanse)

    It's necessary to subclass our base class. (Dataset, Model, Attack, Defense)
    Optional base classes depending on your use case: (ImageSet, ImageFolder, ImageModel)

  2. Register your DIY class in trojanvision

    Example: trojanvision.attacks.class_dict[attack_name]=AttackClass

  3. Create your config files if necessary.
    No need to modify any codes. Just directly add {attack_name}.yml (.json) in the config directory.
  4. Good to go!

Todo List

  1. Sphinx Docs
  2. Unit test

License

TrojanZoo has a GPL-style license, as found in the LICENSE file.

Cite our paper

@InProceedings{pang:2022:eurosp,
      title={TrojanZoo: Towards Unified, Holistic, and Practical Evaluation of Neural Backdoors}, 
      author={Ren Pang and Zheng Zhang and Xiangshan Gao and Zhaohan Xi and Shouling Ji and Peng Cheng and Ting Wang},
      year={2022},
      booktitle={Proceedings of IEEE European Symposium on Security and Privacy (Euro S\&P)},
}

@inproceedings{pang:2020:ccs, 
    title = "{A Tale of Evil Twins: Adversarial Inputs versus Poisoned Models}", 
    author = {Ren Pang and Hua Shen and Xinyang Zhang and Shouling Ji and Yevgeniy Vorobeychik and Xiapu Luo and Alex Liu and Ting Wang}, 
    year = {2020}, 
    booktitle = {Proceedings of ACM SAC Conference on Computer and Communications (CCS)},
}

@inproceedings{pang:2020:kdd, 
    title = "{AdvMind: Inferring Adversary Intent of Black-Box Attacks}", 
    author = {Ren Pang and Xinyang Zhang and Shouling Ji and Xiapu Luo and Ting Wang}, 
    year = {2020}, 
    booktitle = {Proceedings of ACM International Conference on Knowledge Discovery and Data Mining (KDD)},
}