Collaborative Variational Bandwidth Auto-encoder

The codes are associated with the following paper:

Collaborative Variational Bandwidth Auto-encoder for Recommender Systems,
Yaochen Zhu and Zhenzhong Chen.
ArXiv.2105.07597, Preprints. 2021. [pdf].

Environment

The codes are written in Python 3.6.5 with the following packages.

• numpy == 1.16.3
• pandas == 0.21.0
• tensorflow-gpu == 1.15.0
• tensorflow-probability == 0.8.0

Datasets

The processed datasets can be found here.

For usage, create a data folder and move in the unzipped datasets.

Examples to run the codes

To reproduce the comparison results in Table 2:

• Layerwise pretrain the user feature VAE:

  python pretrain_vae.py --dataset Name --split [0-9]

• Iteratively train the collarabotive and feature part of VBAE:

  python train_vbae.py --dataset Name --split [0-9]

• Evaluate the model and summarize the results into a pivot table

  python predict.py --dataset Name --split [0-9]

  python summarize.py

To reproduce the bandwidth analysis results in Table 3:

• Summarize the average, std and corr of bandwidth into the model folder

  python analyse_bandwidth.py --dataset Name --split [0-9]

For more advanced argument usage, run the code with --help argument.

Citation

If you find our codes helpful, please kindly cite the following paper. Thanks!

@article{vbae_zhu2021,
  title={Collaborative Variational Bandwidth Auto-encoder for Recommender Systems},
  author={Zhu, Yaochen and Chen, Zhenzhong},
  booktitle={arXiv preprint arXiv:2105.07597},
  year={2021},
}