Fairness in Music Recommender Systems: A Stakeholder-Centered Mini Review.

Area of Science:

• Computer Science
• Music Information Retrieval
• Algorithmic Fairness

Background:

• Recommender systems significantly influence user experience and artist visibility on music streaming platforms.
• Algorithmic fairness is increasingly demanded due to the potential for biases in automated decision-making processes.
• The music domain has specific fairness challenges due to multiple stakeholders with diverse needs.

Purpose of the Study:

• To review and synthesize existing literature on fairness in music recommender systems.
• To examine fairness from the perspectives of individual stakeholders and their combined interests.
• To identify research gaps and future directions for improving fairness in music recommendations.

Main Methods:

• A narrative literature review approach was employed.
• Analysis focused on fairness considerations across different stakeholders in music recommendation.
• Specific attention was given to gender fairness and popularity bias in existing research.

Main Results:

• Existing research predominantly analyzes the current state of fairness in music recommender systems.
• Studies address gender bias by examining recommendation quality across user groups and artist representation.
• Popularity bias is often discussed from a user viewpoint, with less focus on its impact on artist representation.

Conclusions:

• The majority of studies describe fairness issues rather than proposing solutions.
• There is a significant need for research focused on developing and implementing fairness-improving approaches.
• Future research should prioritize creating more equitable music recommender systems.