Mind Individual Information! Principal Graph Learning for Multimedia Recommendation

September 16, 2025 ยท View on GitHub

Introduction

This is the Pytorch implementation for our AAAI 2025 paper [Oral Paper]:

AAAI 2025. Penghang Yu, Zhiyi Tan, Guanming Lu, Bing-Kun Bao(2025). Mind Individual Information! Principal Graph Learning for Multimedia Recommendation

Enviroment Requirement

  • python 3.8
  • Pytorch 1.12
  • cuda 10.2

Dataset

We provide three processed datasets: Baby, Sports and Clothing. Download from Google Drive: Baby/Sports/Clothing

Training

cd ./src
python main.py

Performance Comparison

Citing MGCN

If you find PGL useful in your research, please consider citing our paper.

@article{yu2025pgl,
  title={Mind Individual Information! Principal Graph Learning for Multimedia Recommendation},
  author={Yu, Penghang and Tan, Zhiyi and Lu, Guanming and Bao, Bing-Kun},
  booktitle={Proceedings of the 31st ACM International Conference on Multimedia},
  volume={39},
  number={12},
  journal={Proceedings of the AAAI Conference on Artificial Intelligence},
  year={2025},
  pages={13096-13105}
}

The code is released for academic research use only. For commercial use, please contact Penghang Yu.

Acknowledgement

The code is based on MMRec. Thank for their work.