Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/edisonleeeee/spikegcl
[ICLR 2024] Official implementation of Spiking Graph Contrastive Learning (0️⃣1️⃣ SpikeGCL)
https://github.com/edisonleeeee/spikegcl
graph-contrastive-learning graph-neural-networks iclr spiking-neural-networks
Last synced: about 2 months ago
JSON representation
[ICLR 2024] Official implementation of Spiking Graph Contrastive Learning (0️⃣1️⃣ SpikeGCL)
- Host: GitHub
- URL: https://github.com/edisonleeeee/spikegcl
- Owner: EdisonLeeeee
- Created: 2023-05-13T02:07:01.000Z (over 1 year ago)
- Default Branch: master
- Last Pushed: 2024-05-08T07:11:42.000Z (8 months ago)
- Last Synced: 2024-05-08T08:28:22.006Z (8 months ago)
- Topics: graph-contrastive-learning, graph-neural-networks, iclr, spiking-neural-networks
- Language: Python
- Homepage: https://openreview.net/forum?id=LnLySuf1vp
- Size: 908 KB
- Stars: 12
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
0️⃣1️⃣ SpikeGCL (Spiking Graph Contrastive Learning)
A Graph is Worth 1-bit Spikes: When Graph Contrastive Learning Meets Spiking Neural Networks
Jintang Li1, Huizhe Zhang1, Ruofan Wu2, Zulun Zhu3, Baokun Wang2, Changhua Meng2, Zibin Zheng1, Liang Chen1
1Sun Yat-sen University, 2Ant Group, 3Nanyang Technological University
arXiv ([arXiv:2305.19306](https://arxiv.org/abs/2305.19306)),
OpenReview ([ICLR'24](https://openreview.net/forum?id=LnLySuf1vp))
[Poster](https://iclr.cc/media/PosterPDFs/ICLR%202024/18850.png?t=1713083679.2213657) | [Slides](https://iclr.cc/media/iclr-2024/Slides/18850.pdf)
## Environments
> [!NOTE]
> Higher versions should be also compatible.
+ numpy == 1.23.3
+ torch == 1.8+cu111
+ torch-cluster == 1.6.1
+ torch_geometric == 2.3.0
+ torch-scatter == 2.1.1
+ torch-sparse == 0.6.17
+ CUDA 11.1
+ cuDNN 8.0.5
## Model and Results
SpikeGCL adopts a simple GCL architecture and is comprised of a set of peer GNN encoders and a spiking neuron.
The following tables present the performance & efficiency results for standard node classification tasks on several graph benchmark datasets.
## Reproduction
+ Cora
```
python main.py --dataset Cora --threshold 5e-4 --outs 2 --T 64 --bn --epochs 5
```
+ Citeseer
```
python main.py --dataset Citeseer --threshold 5e-3 --T 32 --bn --epochs 5
```
+ Pubmed
```
python main.py --dataset Pubmed --threshold 5e-2 --bn --T 32 --epochs 50
```
+ Computers
```
python main.py --dataset Computers --threshold 5e-2 --outs 32 --bn --T 25
```
+ Photo
```
python main.py --dataset Photo --threshold 5e-2 --T 15 --bn --outs 8 --epochs 50
```
+ CS
```
python main.py --dataset CS --threshold 5e-1 --outs 32 --T 60 --dropout 0. --bn
```
+ Physics
```
python main.py --dataset Physics --T 25 --outs 16 --margin 1 --threshold 5e-2 --bn
```
+ Ogbn-arXiv
```
python main.py --dataset ogbn-arxiv --T 30 --outs 1 --threshold 5e-2 --no_shuffle --bn --dropout 0.
```
+ Ogbn-MAG
```
python main.py --dataset ogbn-mag --T 8 --outs 8 --hids 64 --threshold 5e-3 --no_shuffle --bn
```
## Citation
If you find this repository useful in your research, please consider giving a star ⭐ and a citation
```bibtex
@inproceedings{spikegcl,
title={A Graph is Worth 1-bit Spikes: When Graph Contrastive Learning Meets Spiking Neural Networks},
author={Jintang Li and Huizhe Zhang and Ruofan Wu and Zulun Zhu and Baokun Wang and Changhua Meng and Zibin Zheng and Liang Chen},
booktitle={ICLR},
year={2024},
url={https://openreview.net/forum?id=LnLySuf1vp}
}
```