Generating Potent Poisons and Backdoors from Scratch with Guided Diffusion

This code is the official PyTroch implementation of the Generating Potent Poisons and Backdoors from Scratch with Guided Diffusion. Our implementation is based on Industrial Scale Data Poisoning via Gradient Matching, PyTorch pretrained Diffusion Models and this awesome repo.

Dependencies

• PyTorch => 1.6.*
• torchvision > 0.5.*
• higher [best to directly clone https://github.com/facebookresearch/higher and use pip install .]
• python-lmdb [only if datasets are supposed to be written to an LMDB]

USAGE

The wrapper for the GDP can be found in brew_poison.py. To craft poisons on CIFAR-10 and ResNet18 with the default values, you can use the following command:

python brew_poison.py --diffusion_base_poisons --filter_max_matching --num_poisons 50

There are a bunch of optional arguments in the forest/options.py. Here are some of them:

• --eps, and --budget : Determine the power of backdoor attack.
• --net : which model to attack on.
• --num_poisons : Number of filtered poisons.
• --base_optim_guidance_3_wt : Guidance strength.
• --base_num_steps : Recurrent steps for base poisons.
• --base_poison_loss_w : Classifer guidance weight.

Citation

Please cite our paper:

@misc{souri2024generating,
      title={Generating Potent Poisons and Backdoors from Scratch with Guided Diffusion}, 
      author={Hossein Souri and Arpit Bansal and Hamid Kazemi and Liam Fowl and Aniruddha Saha and Jonas Geiping and Andrew Gordon Wilson and Rama Chellappa and Tom Goldstein and Micah Goldblum},
      year={2024},
      eprint={2403.16365},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}