Fairness in Music Recommendation

What makes a song a smash hit these days? Beautiful musical innovations? Deeply touching songwriting? Raw talent?

Probably not any of those things. Its more likely that someone was willing to pay to have a recommendation algorithm give their song a boost. The help of an algorithm means more streams, more publicity, and more merch sales.

In order to explore this topic, I am teaming up with Sean McQuillan and professor Doug Turnbull to explore and evaluate state-of-the-art machine learning algoritms. Comparing the preformace of these models to Spotify can illuminate commerical bias.

We currently in the process of researching. We will be submitting a paper of our findings to the annual conference of the International Society for Music Information Retrieval (ISMIR) 2021! So check back in May to see how our research went.