Music recommendation diversity: a tentative framework and preliminary results

Open Access

Abstract

Comunicació presentada al 1st Workshop on Designing Human-Centric MIR Systems celebrat el 2 de novembre de 2019 a Delft, Països Baixos, en el marc del 20th annual conference of the International Society for Music Information Retrieval (ISMIR). ; Music recommendations are increasingly part of the listening experience of people all over the world, especially in the context of streaming services. In this scenario, recommender systems' role is to help users in finding music that can fit their interests and tastes. However, Westerncentric perspectives in systems' design are often subject to criticisms because of their power of reinforcing already existing cultural bias and therefore potentially impacting negatively on the music distribution mechanisms. In our research proposal, we aim to address the problem of assessing the impact of music recommendation diversity, or the lack thereof. This requires 1) the formalization of a working definition of diversity in the music field 2) the development of evaluation practices for estimating diversity in the context of music recommender systems 3) the observation of emerging impact due to music recommendation diversity 4) the proposal of countermeasures for mitigating negative or reinforcing positive impact observed. Basing on already known consequences of information technologies in political, economic and social areas, our goal is to understand the cultural impact that music recommender systems can have on our society. ; This work is partially supported by the European Commission under the TROMPA project (H2020 770376).