/SUBLIME

A PyTorch implementation of "Towards Unsupervised Deep Graph Structure Learning", WWW-22

Primary LanguagePythonMIT LicenseMIT

Towards Unsupervised Deep Graph Structure Learning

This is the source code of WWW-2022 paper "Towards Unsupervised Deep Graph Structure Learning" (SUBLIME).

The proposed framework

REQUIREMENTS

This code requires the following:

  • Python==3.7
  • PyTorch==1.7.1
  • DGL==0.7.1
  • Numpy==1.20.2
  • Scipy==1.6.3
  • Scikit-learn==0.24.2
  • Munkres==1.1.4
  • ogb==1.3.1

USAGE

Step 1: All the scripts are included in the "scripts" folder. Please get into this folder first.

cd scripts

Step 2: Run the experiments you want:

[Cora]Node classification @ structure inference:

bash cora_si.sh

[Cora]Node classification @ structure refinement:

bash cora_sr.sh

[Cora]Node clustering @ structure refinement:

bash cora_clu.sh

[Citeseer]Node classification @ structure inference:

bash citeseer_si.sh

[Citeseer]Node classification @ structure refinement:

bash citeseer_sr.sh

[Citeseer]Node clustering @ structure refinement:

bash citeseer_clu.sh

[Pubmed]Node classification @ structure inference:

bash pubmed_si.sh

[Pubmed]Node classification @ structure refinement:

bash pubmed_sr.sh

Cite

If you compare with, build on, or use aspects of SUBLIME framework, please cite the following:

@inproceedings{liu2022towards,
  title={Towards unsupervised deep graph structure learning},
  author={Liu, Yixin and Zheng, Yu and Zhang, Daokun and Chen, Hongxu and Peng, Hao and Pan, Shirui},
  booktitle={Proceedings of the ACM Web Conference 2022},
  pages={1392--1403},
  year={2022}
}