DPAO

This is our PyTorch implementation of the paper:

Dual Policy Learning for Aggregation Optimization in Graph Neural Network-based Recommender Systems. Heesoo Jung, Sangpil Kim and Hogun Park.
In The Web Conference (WWW) 2023.

The DPAO algorithms adaptively optimizes the number of GNN layers for each user and item by utilizing RL.
Please check the Paper for more details.

Environment Requirement

The code has been tested running under Python 3.9.15. The required packages are:

• pytorch==1.12.0
• dgl-cu116
• numpy==1.23.4
• scikit-learn==1.2.1
• scipy==1.9.3
• pandas==1.5.3

Reproducing Paper's Experiment

KG-based

Amazon-Book

cd KG
python main_DPAO.py --data_name amazon-book --conv_dim_list '[128, 64, 32, 16]' --mess_dropout '[0.1, 0.1, 0.1, 0.1]' --aggregation_type 'gcn' --evaluate_every 1 --lr 0.0001 --n_epoch -1

Non-KG-based

Gowalla

cd NonKG
python main.py --dataset Gowalla

MovieLens1M

cd NonKG
python main.py --dataset ml-1m

Amazon-Book

cd NonKG
python main.py --dataset amazon-book

Citing

If you want to use our codes and papers in your research, please consider citing the paper:

    @inproceedings{DPAO,
    author={Heesoo Jung, Sangpil Kim and Hogun Park},
    title={Dual Policy Learning for Aggregation Opimization in Graph Neural Network-based Recommender Systems.},
    booktitle={Proceedings of the Web Conference},
    year={2023}
    }

Acknowledgement

We refer to the code of KGAT and NGCF. Thanks for their contribution