Shifting Consumption towards Diverse Content on Music Streaming Platforms


Algorithmic recommendations shape music consumption at scale, and understanding the impact of various algorithmic models on how content is consumed is a central question for music streaming platforms. The ability to shift consumption towards less popular content and towards content different from user’s typical historic tastes not only affords the platform ways of handling issues such as filter bubbles and popularity bias, but also contributes to maintaining a healthy and sustainable consumption patterns necessary for overall platform success.

In this work, we view diversity as an enabler for shifting consumption and consider two notions of music diversity, based on taste similarity and popularity, and investigate how four different recommendation approaches optimized for user satisfaction, fare on diversity metrics. To investigate how the ranker complexity influences diversity, we use two well-known rankers and propose two new models of increased complexity: a feedback aware neural ranker and a reinforcement learning (RL) based ranker. We demonstrate that our models lead to gains in satisfaction, but at the cost of diversity.
Such trade-off between model complexity and diversity necessitates the need for explicitly encoding diversity in the modeling process, for which we consider four types of approaches: interleaving based, submodularity based, interpolation, and RL reward modeling based. We find that our reward modeling based RL approach achieves the best trade-off between optimizing the satisfaction metric and surfacing diverse content, thereby enabling consumption shifting at scale. Our findings have implications for the design and deployment of practical approaches for music diversification, which we discuss at length.


September 2023 | RecSys

Accelerating Creator Audience Building through Centralized Exploration

Buket Baran, Guilherme Dinis Junior, Antonina Danylenko, Olayinka S. Folorunso, Gösta Forsum, Maksym Lefarov, Lucas Maystre, Yu Zhao

August 2023 | Interspeech

Lightweight and Efficient Spoken Language Identification of Long-form Audio

Winstead Zhu, Md Iftekhar Tanveer, Yang Janet Liu, Seye Ojumu, Rosie Jones

July 2023 | SIGIR

Hear Me Out: A Study on the Use of the Voice Modality for Crowdsourced Relevance Assessments

Nirmal Roy, Agathe Balayn, David Maxwell, Claudia Hauff.