qslim / PDF

Towards Better Graph Representation Learning with Parameterized Decomposition & Filtering

Towards Better Graph Representation Learning with Parameterized Decomposition & Filtering

This is the code of the paper "Towards Better Graph Representation Learning with Parameterized Decomposition & Filtering".

Requirements

The following packages need to be installed:

• pytorch==1.13.0
• dgl==0.9.1
• ogb==1.3.5
• numpy
• easydict
• tensorboard
• tqdm
• json5

Usage

ZINC

• Change your current directory to zinc;
• Download the dataset: sh script_download_dataset.sh;
• Configure hyper-parameters in ZINC.json;
• Start training: sh run_script.sh.

ogbg-molpcba

• Change your current directory to ogbg/mol;
• Configure hyper-parameters in ogbg-molpcba.json.json);
• Start training: sh run_script.sh.

TUDataset

• Change your current directory to tu;
• configure hyper-parameters in configs/<dataset>.json;
• Set dataset name in run_script.sh;
• Start training: sh run_script.sh.

Reference

@inproceedings{yang2023towards,
  title = {Towards Better Graph Representation Learning with Parameterized Decomposition \& Filtering},
  author = {Mingqi Yang and Wenjie Feng and Yanming Shen and Bryan Hooi},
  booktitle = {Proceedings of the 40th International Conference on Machine Learning},
  year = {2023},
}

License

MIT License