Skip to content

PKU-ML/LaplacianCanonization

Repository files navigation

Laplacian Canonization: A Minimalist Approach to Sign and Basis Invariant Spectral Embedding

Source code for the paper "Laplacian Canonization: A Minimalist Approach to Sign and Basis Invariant Spectral Embedding", NeurIPS 2023.

The implementation of our MAP algorithm is in data/map.py, along with codes for verifying correctness and counting the ratio of assumption violations.

To download the datasets, run data/script_download_ZINC.sh. The OGBG-MOL* datasets are automatically downloaded from OGB. To run the experiments, use the scripts in scripts/.

Attribution: Our code is built on top of the [SignNet repo] by Lim et al. in 2022, which in turn builds off of the setup in [LSPE repo] by Dwivedi et al. in 2021.

If you use our code, please cite

@inproceedings{laplacian-canonization,
    title={{Laplacian Canonization: A Minimalist Approach to Sign and Basis Invariant Spectral Embedding}},
    author={Ma, George and Wang, Yifei and Wang, Yisen},
    booktitle={NeurIPS},
    year={2023}
}

About

Official code for NeurIPS 2023 paper "Laplacian Canonization: A Minimalist Approach to Sign and Basis Invariant Spectral Embedding".

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published