Model-Agnostic Counterfactual Reasoning for Eliminating Popularity Bias in Recommender System

By Tianxin Wei, Fuli Feng, Jiawei Chen, Chufeng Shi, Ziwei Wu, Jinfeng Yi, Xiangnan He. Knowledge Discovery and Data Mining, 2020

403 citations recorded by the source corpus.

The general aim of the recommender system is to provide personalized suggestions to users, which is opposed to suggesting popular items. However, the normal training paradigm, i.e., fitting a recommender model to recover the user behavior data with pointwise or pairwise loss, makes the model biased towards popular items. This results in the terrible Matthew effect, making popular items be more frequently recommended and become even more popular. Existing work addresses this issue with Inverse Propensity Weighting (IPW), which decreases the impact of popular items on the training and increases the impact of long-tail items. Although theoretically sound, IPW methods are highly sensitive to the weighting strategy, which is notoriously difficult to tune. In this work, we explore the popularity bias issue from a novel and fundamental perspective --- cause-effect. We identify that popularity bias lies in the direct effect from the item node to the ranking score, such that an item's intrinsic property is the cause of mistakenly assigning it a higher ranking score. To eliminate popularity bias, it is essential to answer the counterfactual question that what the ranking score would be if the model only uses item property. To this end, we formulate a causal graph to describe the important cause-effect relations in the recommendation process. During training, we perform multi-task learning to achieve the contribution of each cause; during testing, we perform counterfactual inference to remove the effect of item popularity. Remarkably, our solution amends the learning process of recommendation which is agnostic to a wide range of models --- it can be easily implemented in existing methods. We demonstrate it on Matrix Factorization (MF) and LightGCN [20], which are representative of the conventional and SOTA model for collaborative filtering. Experiments on five real-world datasets demonstrate the effectiveness of our method.

Subfields: Artificial Intelligence, Information Systems

View Paper: https://www.semanticscholar.org/paper/8c01e5b581024b967bc09b9eb5ee5424bbbcb5fd

How researchers see this paper

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

In Artificial Intelligence

1. Xiangnan He
2. Xiang Wang
3. Fuli Feng
4. Meng Wang
5. Tat-seng Chua
6. Jiawei Chen
7. Hande Dong
8. Liqiang Nie

In Information Systems

1. Xiangnan He
2. Xiang Wang
3. Fuli Feng
4. Meng Wang
5. Tat-seng Chua
6. Jiawei Chen
7. Hande Dong
8. Liqiang Nie

Citations in the corpus

• Bias and Debias in Recommender System: A Survey and Future Directions (2020, background): “Causal Graph Leverage causal graph to elucidate and mitigate popularity bias Explainable Requiring assumptions on data generation [171, 184, 212, 219, 221]”