Towards Building a Lightweight and Powerful Computation Graph for Scalable GNN

Accepted by WISE 2024 paper link

How to Run

Download dataset from Google Drive link or BaiduYun link(code:f1ao) and put it at correct place:

/home/user/tot_code
│   
└───graphsage
|   
└───graphsaint
| 
└───dataset/
|   |   flickr/
|   |   ogbn-arxiv/
|   |   ...

Only four files are needed: adj_full.npz, class_map.json, feats.npy and role.json. These public datasets are collected by GraphSAINT and are irrelevant to this paper.

If you want to run certain algorithm(e.g. NESC):

cd NESC
bash run.sh

Results will be saved at log directory.

If you find our work useful, please cite our paper:

@InProceedings{NESC, 
author={Gong, Chang and Yang, Boyu and Zheng, Weiguo and Huang, Xing and Huang, Weiyi and Wen, Jie and Sun, Shijie and Yang, Bohua}, 
title={Towards Building a Lightweight and Powerful Computation Graph for Scalable GNN}, 
booktitle={Web Information Systems Engineering -- WISE 2024}, 
year={2025}, 
pages={177--192} 
}