Awesome

CIKM2023DiRec

This is the official code for CIKM 2023 Long Paper:

Dual Intents Graph Modeling for User-centric Group Discovery

Our paper has been released on arXiv now (Paper Link)！

Other Mateirals: Pre-recorded Video Full Paper

overview

Cite

If you make advantage of DiRec in your research, please cite the following in your manuscript:

@inproceedings{wu2023direc,
  title={Dual Intents Graph Modeling for User-centric Group Discovery},
  author={Wu, Xixi and Xiong, Yun and Zhang, Yao and Jiao, Yizhu and Zhang, Jiawei},
  booktitle={Proceedings of the 32nd ACM International Conference on Information and Knowledge Management(CIKM)},
  year={2023},
  organization={ACM}
}

Contents in This Repo

In DiRec folder, we provide the implementation of DiRec model.

In data folder, we provide all three experimental datasets. For more details, you can see that folder's README file.

In baselines folder, we also release our implementations of all baseline models. We consider three types of baseline models: Group Recommendation (GR) models, Recommender Models (RS), and User-centric Group Discovery (UGD) models. We provide the implementation of all 9 baselines compared by our paper.

├── DiRec                   (DiRec Implementation)
│   ├── dataloader.py
│   ├── datautil.py
│   ├── main.py
│   ├── metrics.py
│   ├── model.py
│   └── run.sh
├── README.md
├── baselines
│   ├── CFAG                (UGD baseline)
│   ├── GRModels            (GR baselines, AGREE and ConsRec)
│   │   ├── AGREE
│   │   ├── ConsRec
│   └── RSModels            (RS baselines, including MF-BPR, NFCG, LightGCN, SGL, SimGCL, DCCF)
├── data
│   ├── Mafengwo
│   ├── README.md
│   ├── Steam
│   └── Weeplace
└── figs

Dependencies

Python3
PyTorch 1.13.0
PyTorch-Geometric 2.3.1

Note that it may need to appropriately install the package torch-geometric based on the CUDA version (or CPU version if GPU is not available). Please refer to the official website https://pytorch-geometric.readthedocs.io/en/latest/notes/installation.html for more information of installing prerequisites.

scipy 1.10.1

Reproducibility

Use run.sh in DiRec/baseline folders to run the codes and reproduce the published results.

Performance