thunlp / AGE

Source code and dataset for KDD 2020 paper "Adaptive Graph Encoder for Attributed Graph Embedding"

AGE

Source code and datasets for KDD 2020 paper "Adaptive Graph Encoder for Attributed Graph Embedding"

---

Requirements

Please make sure your environment includes:

python (tested on 3.7.4)
pytorch (tested on 1.2.1)

Then, run the command:

pip install -r requirements.txt

Run

Run AGE on Cora dataset:

python train.py --dataset cora --gnnlayers 8 --upth_st 0.011 --lowth_st 0.1 --upth_ed 0.001 --lowth_ed 0.5

To reproduce the node clustering experiment results, please follow our hyper-parameter settings:

Dataset	gnnlayers	upth_st	upth_ed	lowth_st	lowth_ed
Cora	8	0.0110	0.0010	0.1	0.5
Citeseer	3	0.0015	0.0010	0.1	0.5
Wiki	1	0.0011	0.0010	0.1	0.5
Pubmed	35	0.0013	0.0010	0.7	0.8

For link prediction, please run link_pred.py. We did not tune hyper-parameters for link prediction, so you can tune all kinds of hyper-parameters to get better performance.

Cite

If you use the code, please cite our paper:

@inproceedings{cui2020adaptive,
  title={Adaptive Graph Encoder for Attributed Graph Embedding},
  author={Cui, Ganqu and Zhou, Jie and Yang, Cheng and Liu, Zhiyuan},
  booktitle={Proceedings of SIGKDD 2020},
  year={2020}
}