https://github.com/chandlerbang/selftask-gnn

Implementation of paper "Self-supervised Learning on Graphs:Deep Insights and New Directions"
https://github.com/chandlerbang/selftask-gnn

deep-learning graph-mining graph-neural-networks machine-learning self-supervised-learning

Last synced: about 2 months ago
JSON representation

Implementation of paper "Self-supervised Learning on Graphs:Deep Insights and New Directions"

• Host: GitHub
• URL: https://github.com/chandlerbang/selftask-gnn
• Owner: ChandlerBang
• Created: 2020-06-17T17:40:13.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2022-06-22T02:20:25.000Z (over 2 years ago)
• Last Synced: 2023-03-04T17:22:49.375Z (almost 2 years ago)
• Topics: deep-learning, graph-mining, graph-neural-networks, machine-learning, self-supervised-learning
• Language: Python
• Homepage: https://arxiv.org/abs/2006.10141
• Size: 4.98 MB
• Stars: 86
• Watchers: 2
• Forks: 20
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# SelfTask-GNN

A PyTorch implementation of "Self-supervised Learning on Graphs: Deep Insights and New Directions". [[paper]](https://arxiv.org/abs/2006.10141)  

In this paper, we first deepen our understandings on when, why, and which strategies of SSL work with GNNs by empirically studying numerous basic SSL pretext tasks on graphs. Inspired by deep insights from the empirical studies, we propose a new direction *SelfTask* to build advanced pretext tasks that are able to achieve state-of-the-art performance on various real-world datasets.

## Requirements

See that in https://github.com/ChandlerBang/SelfTask-GNN/blob/master/requirements.txt

## Run our code

Clone the repository

```

git clone https://github.com/ChandlerBang/SelfTask-GNN.git

cd SelfTask-GNN

pip install -r requirements.txt

```

You need to further install `ica` package

```

pip uninstall ica # in case you have installed it before

git clone https://github.com/ChandlerBang/ica.git 

cd ica

python setup.py install

```

To reproduce the performance reported in the paper, you can run the bash files in folder `scripts`. 

```

sh scripts/selftask/cora_CorrectedLabel_ICA.sh

sh scripts/selftask/cora_CorrectedLabel_LP.sh

```

## Acknowledgement

This repository is modified from DropEdge [(https://github.com/DropEdge/DropEdge)](https://github.com/DropEdge/DropEdge). We sincerely thank them for their contributions.

## Cite

For more information, you can take a look at the [paper](https://arxiv.org/abs/2006.10141) 

If you find this repo to be useful, please cite our paper. Thank you.

```

@misc{jin2020selfsupervised,

    title={Self-supervised Learning on Graphs: Deep Insights and New Direction},

    author={Wei Jin and Tyler Derr and Haochen Liu and Yiqi Wang and Suhang Wang and Zitao Liu and Jiliang Tang},

    year={2020},

    eprint={2006.10141},

    archivePrefix={arXiv},

    primaryClass={cs.LG}

}

```