Heterogeneous Graph Contrastive Learning with Meta-path Contexts and Adaptively Weighted Negative Samples

MEOW & AdaMEOW

This is the source code of two papers:

• $\textbf{MEOW}:$ "Heterogeneous Graph Contrastive Learning with Meta-path Contexts and Weighted Negative Samples (SDM 2023)"

• $\textbf{AdaMEOW}:$ "Heterogeneous Graph Contrastive Learning with Meta-path Contexts and Adaptively Weighted Negative Samples (TKDE 2024)".

Environment Settings

python==3.9.0 \ scipy==1.8.1 \ torch==1.12.0 \ numpy==1.23.0 \ scikit_learn==1.1.1\ faiss-gpu==1.7.2

Usage

Go into ./code_meow, and then you can use the following commend to run our model MEOW;
or go into ./code_adameow, and then you can use the following commend to run our model AdaMEOW:

python main.py acm --gpu=0

Here, "acm" can be replaced by "dblp", "aminer","imdb".

Some files in the './data' could not be uploaded because they were over 25MB. All the data files we store in url：https://pan.baidu.com/s/1vlBrC4S7EZgowGyHGF8apg pwd：n84e

Citation

@article{yu2024heterogeneous,
  title={Heterogeneous Graph Contrastive Learning With Meta-Path Contexts and Adaptively Weighted Negative Samples},
  author={Yu, Jianxiang and Ge, Qingqing and Li, Xiang and Zhou, Aoying},
  journal={IEEE Transactions on Knowledge and Data Engineering},
  year={2024},
  publisher={IEEE}
}