Modular Debiasing of Latent User Representations in Prototype-based Recommender Systems @ ECML-PKDD'24

This repository hosts the code and and the settings for the paper "Modular Debiasing of Latent User Representations in Prototype-based Recommender Systems" by Alessandro B. Melchiorre, Shahed Masoudian, Deepak Kumar, and Markus Schedl at ECML-PKDD'24.

Installation

Environment

• Install the environment with conda env create -f modprotodebias.yml
• Activate the environment with conda activate modprotodebias

Data

• move into the folder with cd data/<dataset_folder>
• run python <dataset_name>_processor.py

If you have problems with the LFM2b data, ping me and I'll be happy to help

Pre-Trained Models

• download the pre-trained ProtoMF models from here
• place the two folders inside pre_trained_models folder (default)
• (optional) adjust the path files in the conf.yml if you have issues

Usage

Adjust the configuration of your experiment in run_full_debiasing.py.

The experiments can be started with

python start.py run_full_debiasing

or define sweep configurations to use with the wandb sweep command

wandb sweep sweep_config.yaml

Cite

@inproceedings{melchiorre2024modular,
  title = {Modular Debiasing of Latent User Representations in Prototype-based Recommender Systems},
  author = {Melchiorre, Alessandro B. and Masoudian, Shahed and Kumar, Deepak and Schedl, Markus},
  booktitle = {Proceedings of 2024 Joint European Conference on Machine Learning and Knowledge Discovery in Databases (ECML PKDD)},
  year = {2024},
}

License

The code in this repository is licensed under the MIT License. For details, please see the LICENSE file.

Acknowledgments

This research was funded in whole or in part by the Austrian Science Fund (FWF): P36413, P33526, and DFH-23, and by the State of Upper Austria and the Federal Ministry of Education, Science, and Research, through grant LIT-2021-YOU-215.