Positional Encoding meets Persistent Homology on Graphs (arxiv)

Yogesh Verma | Amauri H. Souza | Vikas Garg

The repository is developed on the intersection of RePHINE, and SPE. Please take a look at their repos for specific requirements.

Training

To run the method on ZINC or Alchemy, do the following:

cd SPE/zinc/ or SPE/alchemy
python -u runner.py --config_dirpath ../configs/alchemy (or zinc) --config_name SPE_gine_gin_mlp_pe37.yaml --seed 0

One can change the type of PE method in the config files.

Citation

If you find this repository useful in your research, please consider citing the following paper:

@inproceedings{
pipe,
title={Positional Encoding meets Persistent Homology on Graphs},
author={Verma, Yogesh and Souza, Amauri H and Garg, Vikas},
booktitle={The Forty Second International Conference on Machine Learning},
year={2025}
}