Disclamer: We are still working on improving Niseko and lots of changes are expected to be made, please stay tuned!

Niseko is a large-scale meta-learning platform with a high volume of pipeline traces across hundreds of datasets and easy-to-use API for analytics, evaluation and deployment.

Please install the dependencies through

pip install -r requirements.txt

Please download the data here.

All the notebooks are under examples/notebooks.

If you have any questions about Niseko, please reach out to [email protected].

If you find the idea or code useful for your research, please cite our paper:

@inproceedings{shang2019metalearn,
  title={ME-Net: Towards Effective Adversarial Robustness with Matrix Estimation},
  author={Shang, Zeyuan and Zgraggen, Emanuel and Eichmann, Philipp and Kraska, Tim},
  booktitle={Workshop on Meta-Learning at NeurIPS 2019},
  year={2019},
  url={http://metalearning.ml/2019/papers/metalearn2019-shang.pdf},
}