Rearchitecting Classification Frameworks For Increased Robustness (YOPO Fork)

YOPO adversarial training experiments: Brian Tang and Kassem Fawaz

Varun Chandrasekaran, Brian Tang, Nicholas Papernot, Kassem Fawaz, Somesh Jha, Xi Wu

Sample code for our paper, "Rearchitecting Classification Frameworks For Increased Robustness" arXiv

Requirements

This code is tested with Python 3.5.2 Other required packages can be found in requirements.txt Sample virtual environment commands:

python3 -m venv path_to_environment/
source path_to_environment/bin/activate
pip install -r requirements.txt

How to run our code

• train.py -- Train a YOPO model
• config.py -- Contains hyperparameters
• dataset.py -- Data processing and equivalence classes
• eval.py -- Evaluates a model checkpoint
• eval_test_ids.py -- Does end to end evaluation of the hierarchy
• loss.py -- Contains loss function
• training_function.py -- Contains training functionality

You can change all the hyper-parameters in config.py. And change network in network.py Actually code in above mentioned director is very flexible and can be easiliy modified. It can be used as a template.

YOPO training

Go to directory experiments/CIFAR10/binary run python train.py -d <which_gpu> change checkpoint filename and run python eval_test_ids.py

Go to directory experiments/CIFAR10/4leaf run python train.py -d <which_gpu> change checkpoint filename and run python eval_test_ids.py

Go to directory experiments/CIFAR10/6leaf run python train.py -d <which_gpu> change checkpoint filename and run python eval_test_ids.py

Go to directory experiments/CIFAR10/full run python train.py -d <which_gpu> run python eval.py --resume <path_to_checkpoint>

You can change all the hyper-parameters in config.py. And change network in network.py Runing this code for the first time will dowload the dataset in ./experiments/CIFAR10/data/, you can modify the path in dataset.py

Miscellaneous

• Forked from this repository
• We leave integrating other datasets and generalizable code for creating and evaluating hierarchies as future work.
• Please open an issue or contact Brian Tang (byron123t@gmail.com) if there is any question.

Cite

@article{chandrasekaran2019,
  title={Rearchitecting Classification Frameworks For Increased Robustness},
  author={Chandrasekaran, Varun and Tang, Brian and Papernot, Nicolas and Fawaz, Kassem and Jha, Somesh and Wu, Xi},
  journal={arXiv preprint arXiv:1905.10900},
  year={2019}
}