Barlow Twins Guided Filter(BTGF)

This is code for Upper Bounding Barlow Twins: A Novel Filter for Multi-relational Clustering AAAI-24.

Overall node clustering result:

Datasets

The statistics of the datasets are as follows:

DBLP and Amazon can be found on Google Drive.

Usage

The learning rate and weight decay of the optimizer are set to $1e^{−2} and \1e^{−3}$.

The filter’s parameters $k$ and $\alpha$ are tuned in $[1, 2, 3, 4]$ and $[1, 10, 100, 1000]$, respectively.

You can run BTGF with commands in the script.sh

python main.py -dataset ACM -epoch 400 -lr 1e-2 -wd 1e-3 -k 4 -a 10

python main.py -dataset amazon -epoch 400 -lr 1e-2 -wd 1e-3 -k 2 -a 1

python main.py -dataset aminer -epoch 400 -lr 1e-2 -wd 1e-3 -k 3 -a 100

python main.py -dataset DBLP_L -epoch 400 -lr 1e-2 -wd 1e-3 -k 2 -a 1000

BibTex

Please cite our paper if you found our datasets or code helpful.

@inproceedings{qian2024upper,
  title={Upper Bounding Barlow Twins: A Novel Filter for Multi-Relational Clustering},
  author={Qian, Xiaowei and Li, Bingheng and Kang, Zhao},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={38},
  number={13},
  pages={14660--14668},
  year={2024}
}