Source code for the paper "Initializing Then Refining: A Simple Graph Attribute Imputation Network"
W. Tu, S. Zhou, X. Liu, Y. Liu, Z. Cai, E. Zhu, C. Zhang, and J. Cheng.
Accepted by IJCAI 2022. [Paper]
Clone this repo.
git clone https://github.com/WxTu/ITR.git- Python 3.7.11
- Pytorch (1.9.0)
- Numpy 1.25.1
- Sklearn 1.0.2
- Torchvision 0.10.0
- Matplotlib 3.5.1
We adopt four datasets in total, including Cora, Citeseer, Amazon Computer, and Amazon Photo. To train ITR on these datasets, please download them from Baidu Cloud (access code: 1234).
Here we provide an implementation of Initializing Then Refining (ITR) in PyTorch, along with an execution example on Cora (or Citeseer) dataset (due to file size limit). The repository is organised as follows:
ITR.py: defines the architecture of the whole network.utils.py: defines some functions about data processing, evaluation metrics, and others.main.py: the entry point for training and testing.test_X.pyandtest_AX.py: about downstream tasks.
Finally, main.py puts all of the above together and may be used to execute a full training run on Cora (or Citeseer).
Any discussions or concerns are welcomed!
If you use this code for your research, please cite our paper.
@inproceedings{2022ITR,
title={Initializing Then Refining: A Simple Graph Attribute Imputation Network},
author={Wenxuan Tu and Sihang Zhou and Xinwang Liu and Yue Liu and Zhiping Cai and En Zhu and Changwang Zhang and Jieren Cheng},
booktitle={Proceedings of The Thirty-First International Joint Conference on Artificial Intelligence},
pages={3494-3500},
year={2022}
}
All rights reserved. Licensed under the Apache License 2.0.
The code is released for academic research use only. For commercial use, please contact [wenxuantu@163.com].
X. Chen, S. Chen, J. Yao, et al. Learning on Attribute-Missing Graphs. IEEE TPAMI, 2022.
--https://github.com/xuChenSJTU/SAT-master-online