https://github.com/jonasrauber/analysis-by-synthesis
Analysis by Synthesis — reimplemented
https://github.com/jonasrauber/analysis-by-synthesis
adversarial-examples machine-learning pytorch
Last synced: 10 months ago
JSON representation
Analysis by Synthesis — reimplemented
- Host: GitHub
- URL: https://github.com/jonasrauber/analysis-by-synthesis
- Owner: jonasrauber
- Created: 2019-10-29T21:40:22.000Z (over 6 years ago)
- Default Branch: public
- Last Pushed: 2020-04-23T10:56:28.000Z (about 6 years ago)
- Last Synced: 2025-04-04T11:12:50.867Z (about 1 year ago)
- Topics: adversarial-examples, machine-learning, pytorch
- Language: Python
- Homepage:
- Size: 3.28 MB
- Stars: 5
- Watchers: 2
- Forks: 5
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Analysis by Synthesis
This is a complete reimplementation of our [Analysis by Synthesis](https://github.com/bethgelab/AnalysisBySynthesis) model.
The experiments in the paper were done with the original implementation. To the best of my knowledge, both are equivalent
and I tried to carefully reproduce all results but I cannot make any guarantees.
## Results from the paper
Our paper can be found on [arXiv](https://arxiv.org/abs/1805.09190). It was accepted at ICLR 2019.
All data from the accuracy-distortion curves can be downloaded in raw form:
* [figures/L2_accuracy_distortion_curves.pickle](figures/L2_accuracy_distortion_curves.pickle)
* [figures/Linf_accuracy_distortion_curves.pickle](figures/Linf_accuracy_distortion_curves.pickle)
* [figures/L0_accuracy_distortion_curves.pickle](figures/L0_accuracy_distortion_curves.pickle)
## BibTex
```bibtex
@inproceedings{schott2018towards,
title={Towards the first adversarially robust neural network model on {MNIST}},
author={Lukas Schott and Jonas Rauber and Matthias Bethge and Wieland Brendel},
booktitle={International Conference on Learning Representations},
year={2019},
url={https://openreview.net/forum?id=S1EHOsC9tX},
}
```