Skip to content

NIRM: Dismantling Complex Networks by a Neural Model Trained from Tiny Networks

Notifications You must be signed in to change notification settings

JiazhengZhang/NIRM

Repository files navigation

NIRM

The official PyTorch implementation of Neural Influence Ranking Model (NIRM) in the following paper:

Jiazheng Zhang, Bang Wang. 2022. Dismantling Complex Networks by a Neural Model 
Trained from Tiny Networks. In CIKM'22, October 17-22, 2022, Atlanta, USA, 10 pages.

Dependencies

  • torch 1.7.1
  • torch-geometric 1.7.2
  • torch-sparse 0.6.9
  • torch-scatter 2.0.7
  • sklearn 0.24.2
  • numpy 1.19.1
  • pandas 1.3.0
  • networkx 2.6.2
  • scipy 1.7.0

Install all dependencies using

pip install -r requirements.txt

Usage

  1. Generate synthetic training dataset:
python GenerateTrainData.py
  1. Modify hyper-parameters in Train.py, and run the following to train the model:
python Train.py
  1. Test the well-trained model on the real-world networks:
python Test.py

we provide a well-trained model for one-pass dismantling in the fold './checkpoints/'.

Citation

Please cite our work if you find our code/paper is helpful to your work.

@inproceedings{zhang2022NIRM,
  title={Dismantling Complex Networks by a Neural Model Trained from Tiny Networks},
  author={Zhang, Jiazheng and Wang, Bang},
  booktitle={Proceedings of the 31st ACM International Conference on Information and Knowledge Management},
  series={CIKM'22},
  year={2022},
  location={Atlanta, Georgia, USA},
  numpages={10}
}

About

NIRM: Dismantling Complex Networks by a Neural Model Trained from Tiny Networks

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages