Metric learning based music recommendation system

Recommender systems use the field of artificial intelligence extensively. Specifically, music recommender systems have experienced a huge expansion, thanks to the emergence of successful online streaming services, like Spotify, Apple Music, or Pandora. As this area of recommendations grows, so do the expectations of users who want to explore the vast amounts of music available at their fingertips. Music recommender systems are faced with many challenges - aspiring to provide better and more personalized recommendations requires to go beyond simple content-based recommendations and tune into listeners’ needs and preferences. This work describes preparation of an engineering thesis being a music recommendation system which uses the Metric Learning approach. It is an attempt to implement a system which delves more into personalized recommendations for users’ listening sessions.