Feedback Loop and Bias Amplification in Recommender Systems

By Masoud Mansoury, Himan Abdollahpouri, Mykola Pechenizkiy, Bamshad Mobasher, Robin D. Burke. International Conference on Information and Knowledge Management, 2020

327 citations recorded by the source corpus.

Recommendation algorithms are known to suffer from popularity bias; a few popular items are recommended frequently while the majority of other items are ignored. These recommendations are then consumed by the users, their reaction will be logged and added to the system: what is generally known as a feedback loop. In this paper, we propose a method for simulating the users interaction with the recommenders in an offline setting and study the impact of feedback loop on the popularity bias amplification of several recommendation algorithms. We then show how this bias amplification leads to several other problems such as declining the aggregate diversity, shifting the representation of users' taste over time and also homogenization of the users. In particular, we show that the impact of feedback loop is generally stronger for the users who belong to the minority group.

Subfields: Information Systems, Management Science and Operations Research, Marketing

View Paper: https://www.semanticscholar.org/paper/3fb75b4ada6eb05fcc8d0716755180e0aebcd076

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Systems

1. Xiangnan He
2. Xiang Wang
3. Fuli Feng
4. Meng Wang
5. Tat-seng Chua
6. Yi Tay
7. Jiawei Chen
8. Tianxin Wei

In Management Science and Operations Research

1. Xiangnan He
2. Xiang Wang
3. Meng Wang
4. Fuli Feng
5. Ed H. Chi
6. Lichan Hong
7. Shashank Rajput
8. Ricardo A. Baeza-Yates

In Marketing

1. Yi Tay
2. Jiawei Chen
3. Yongdong Zhang
4. Hande Dong
5. Xiang Wang
6. Fuli Feng
7. Lichan Hong
8. Ed H. Chi

Citations in the corpus

• Measuring Bias (2023, background): “Consequently, the user will see and interact with the future recommendations and as a result, a closed feedback loop happens [23].”
• Bias and Debias in Recommender System: A Survey and Future Directions (2020, background): “Many researchers also study the impact of feedback loop on the popularity bias [57], [58], [59].”
• Causal Intervention for Leveraging Popularity Bias in Recommendation (2021, background): “On one hand, the user-item interaction data usually exhibits long-tail distribution on item popularity — a few head items occupy most of the interactions whereas the majority of items receive relatively little attention [2, 30].”
• Recommender Systems with Generative Retrieval (2023, methodology): “Instead of traditional query-candidate matching approaches, our method is uses an end-to-end”