Cross-Dataset Propensity Estimation for Debiasing Recommender Systems

Abstract

Datasets for training recommender systems are often subject to distribution shift induced by users' and recommenders' selection biases. In this paper, we study the impact of selection bias on datasets with different quantization. We then leverage two differently quantized datasets from different source distributions to mitigate distribution shift by applying the inverse probability scoring method from causal inference. Empirically, our approach gains significant performance improvement over single-dataset methods and alternative ways of combining two datasets.

PDF NeurIPSW OpenReview Semantic Scholar

Cite

Text

Li and Dean. "Cross-Dataset Propensity Estimation for Debiasing Recommender Systems." NeurIPS 2022 Workshops: DistShift, 2022.

Markdown

[Li and Dean. "Cross-Dataset Propensity Estimation for Debiasing Recommender Systems." NeurIPS 2022 Workshops: DistShift, 2022.](https://mlanthology.org/neuripsw/2022/li2022neuripsw-crossdataset/)

BibTeX

@inproceedings{li2022neuripsw-crossdataset,
  title     = {{Cross-Dataset Propensity Estimation for Debiasing Recommender Systems}},
  author    = {Li, Fengyu and Dean, Sarah},
  booktitle = {NeurIPS 2022 Workshops: DistShift},
  year      = {2022},
  url       = {https://mlanthology.org/neuripsw/2022/li2022neuripsw-crossdataset/}
}