Algorithmic Collective Action in Recommender Systems: Promoting Songs by Reordering Playlists

2024-03-19

arXiv.org

AI Summary

Long-tail artists suffer from insufficient exposure in intelligent music recommendation systems. Method: We propose a lightweight, non-adversarial playlist co-optimization framework wherein fans strategically insert target niche tracks into their personal playlists. Leveraging the positional sensitivity of Transformer-based recommenders and the inherent long-tail distribution of music consumption, this approach requires minimal intervention—less than 0.01% of training data—without modifying model architecture or retraining. The method relies solely on position-aware insertion policies and statistical modeling. Contribution/Results: Experiments demonstrate up to a 40× increase in recommendation frequency for targeted tracks, with no statistically significant degradation in overall recommendation performance. Newly generated exposure is uniformly distributed across multiple long-tail artists. This work introduces a decentralized, interpretable paradigm for enhancing recommendation fairness—achieving equitable exposure through user-driven, model-agnostic interventions.

Technology Category

• Data Mining & Knowledge Management: Recommender Systems
• Machine Learning: Learning Preferences or Rankings
• Multiagent Systems: Adversarial Agents

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Economics and fairness of platforms and recommendation systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

We investigate algorithmic collective action in transformer-based recommender systems. Our use case is a music streaming platform where a collective of fans aims to promote the visibility of an underrepresented artist by strategically placing one of their songs in the existing playlists they control. We introduce two easily implementable strategies to select the position at which to insert the song with the goal to boost recommendations at test time. The strategies exploit statistical properties of the learner by targeting discontinuities in the recommendations, and leveraging the long-tail nature of song distributions. We evaluate the efficacy of our strategies using a publicly available recommender system model released by a major music streaming platform. Our findings reveal that through strategic placement even small collectives (controlling less than 0.01% of the training data) can achieve up to $40 imes$ more test time recommendations than an average song with the same number of training set occurrences. Focusing on the externalities of the strategy, we find that the recommendations of other songs are largely preserved, and the newly gained recommendations are distributed across various artists. Together, our findings demonstrate how carefully designed collective action strategies can be effective while not necessarily being adversarial.

Problem

• Music Recommendation Systems
• Playlist Optimization
• Undiscovered Artists Promotion

Innovation

• Playlist Optimization
• Music Recommendation Systems
• Song Popularity Dynamics