Consumer-side fairness in recommender systems: a systematic survey of methods and evaluation

Bjørnar Vassøy,Helge Langseth
DOI: https://doi.org/10.1007/s10462-023-10663-5
IF: 9.588
2024-03-30
Artificial Intelligence Review
Abstract:In the current landscape of ever-increasing levels of digitalization, we are facing major challenges pertaining to data volume. Recommender systems have become irreplaceable both for helping users navigate the increasing amounts of data and, conversely, aiding providers in marketing products to interested users. Data-driven models are susceptible to data bias, materializing in the bias influencing the models' decision-making. For recommender systems, such issues are well exemplified by occupation recommendation, where biases in historical data may lead to recommender systems relating one gender to lower wages or to the propagation of stereotypes. In particular, consumer-side fairness, which focuses on mitigating discrimination experienced by users of recommender systems, has seen a vast number of diverse approaches. The approaches are further diversified through differing ideas on what constitutes fair and, conversely, discriminatory recommendations. This survey serves as a systematic overview and discussion of the current research on consumer-side fairness in recommender systems. To that end, a novel taxonomy based on high-level fairness definitions is proposed and used to categorize the research and the proposed fairness evaluation metrics. Finally, we highlight some suggestions for the future direction of the field.
computer science, artificial intelligence
What problem does this paper attempt to address?
### What problem does the paper attempt to solve? The paper primarily explores the issue of consumer-side fairness in recommendation systems and provides a systematic review of it. Specifically: 1. **Background and Challenges**: - With the increasing degree of digitalization and the rapid growth of data, recommendation systems have become indispensable in helping users filter information and assisting providers in marketing products. - Data-driven models are prone to data biases, which are particularly evident in recommendation systems, such as gender wage gaps in job recommendations or the propagation of stereotypes. 2. **Research Focus**: - Consumer-side fairness focuses on reducing the discrimination experienced by users of recommendation systems. The paper points out that there are currently various methods to address this issue, but these methods have different understandings of what constitutes fairness. - By systematically reviewing and discussing existing literature, the paper proposes a new classification system that categorizes related research and proposed fairness evaluation metrics based on high-level definitions of fairness. 3. **Contributions**: - The paper proposes a classification method to categorize consumer-side fairness methods based on high-level conceptual definitions of fairness and how fairness is integrated into recommendation systems. - It provides a comprehensive overview, classification, and comparison of existing consumer-side fairness methods and their fairness evaluations. Through the above work, the paper aims to provide a more systematic understanding and future direction suggestions for fairness research in the field of recommendation systems.