This is the source code for the KDD'22 paper:

Graph Rationalization with Environment-based Augmentations

by Gang Liu (gliu7@nd.edu), Tong Zhao, Jiaxin Xu, Tengfei Luo, Meng Jiang

This code package was developed and tested with Python 3.9.9 and PyTorch 1.10.1. All dependencies specified in the requirements.txt file. The packages can be installed by

pip install -r requirements.txt

Following are the commands to run experiments on polymer or molecule datasets using default settings.

# OGBG-HIV for example
python main_pyg.py --dataset ogbg-molhiv --by_default
# Polymer Oxygen Permeability
python main_pyg.py --dataset plym-o2_prop --by_default

We provide the oxygen permeability dataset (.csv) for polymer graph regression. It can be found in the data/'name'/raw folder.

Update March 26, 2025: We delegated the polymer datasets for GlassTemp, MeltingTemp, and PolyDensit as requested by the NIMS Materials Database, MatNavi.

Binary classification tasks for the OGBG dataset (i.e., HIV, ToxCast, Tox21, BBBP, BACE, ClinTox and SIDER) can be directedly implemented using commands such as --dataset ogbg-molhiv following the instructions of the official OGBG dataset implementations.

If you find this repository useful in your research, please cite our paper:

@inproceedings{liu2022graph,
  title={Graph Rationalization with Environment-based Augmentations},
  author={Liu, Gang and Zhao, Tong and Xu, Jiaxin and Luo, Tengfei and Jiang, Meng},
  booktitle = {Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining},
  publisher = {Association for Computing Machinery},
  pages = {1069–1078},
  numpages = {10},
  year={2022}
}