Official PyTorch implementation of SGRL in 'Exploitation of a Latent Mechanism in Graph Contrastive Learning: Representation Scattering' (NeurIPS 2024).
python==3.9.7
scikit-learn==1.4.2
scipy==1.13.0
networkx==3.2
numpy==1.26.4
torch==2.1.0
torch_geometric==2.5.3
tqdm==4.66.2
More details can be found in env_description.txt.
You can reproduce the results in the paper easily by running the command bash run_{dataset_name}.
Please cite our paper if SGRL helps your work.
@article{he2024exploitation,
title={Exploitation of a Latent Mechanism in Graph Contrastive Learning: Representation Scattering},
author={Dongxiao He, Lianze Shan, Jitao Zhao, Hengrui Zhang, Zhen Wang, Weixiong Zhang},
journal={NeurIPS},
year={2024}
}