wellido/Aries

Source Code of Aries

Aries: Efficient Testing of Deep Neural Networks via Labeling-Free Accuracy Estimation

Requirements

- Tensorflow 2.3
- Keras 2.4.3

project structure

├── baselines                     # code for baselines, CES, PACE, Meta-set, Prediction-based, and Random
├── models                        # how to use dataset and prepare models            
├── results   
    ├── figures_for_rq2           # all figures of RQ2
├── utils                         # Tiny-ImageNet generator
├── Aries.py                      # Aries source code
├── example.py                    # example of using Aries

Others

• PACE code is from https://github.com/pace2019/pace
• CES code is from https://github.com/Lizn-zn/DNNOpAcc
• Meta-set code is from https://github.com/Simon4Yan/Meta-set
• CIFAR10-C data: https://zenodo.org/record/2535967#.YfRacRNKhQI
• Tiny-ImageNet-C data: https://zenodo.org/record/2469796#.YnPl6hMzZhE

Citation

If you use the code in your research, please cite:

    @article{hu2022efficient,
    title={Aries: Efficient Testing of Deep Neural Networks via Labeling-Free Accuracy Estimation},
    author={Hu, Qiang and Guo, Yuejun and Xie, Xiaofei and Cordy, Maxime and Ma, Lei and Papadakis, Mike and Traon, Yves Le},
    journal={arXiv preprint arXiv:2207.10942},
    year={2022}
}