Skip to content

This is the official repository for our paper KAGNNs: Kolmogorov-Arnold Networks meet Graph Learning.

Notifications You must be signed in to change notification settings

RomanBresson/KAGNN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

KAGNNs -- Graph Neural Networks that use Kolmogorov Arnold Networks as their building blocks

This is the official repository for our paper KAGNNs: Kolmogorov-Arnold Networks meet Graph Learning.

All experiments were run with python>=3.11.

The code is split by learning task (among node classification, graph classification, graph regression). Scripts are provided so experiments can be reproduced.

For our KAN-based models, we used the efficient-kan implementation.

Please cite our work if you use code from this repository:

@misc{bresson2024kagnn,
      title={KAGNNs: Kolmogorov-Arnold Networks meet Graph Learning}, 
      author={Roman Bresson and Giannis Nikolentzos and George Panagopoulos and Michail Chatzianastasis and Jun Pang and Michalis Vazirgiannis},
      year={2024},
      eprint={2406.18380},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2406.18380}, 
}

About

This is the official repository for our paper KAGNNs: Kolmogorov-Arnold Networks meet Graph Learning.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published