Using Shared Vector Representations of Words and Chords in Music for Genre Classification

With so much music readily available for consumption today, it has never been more important to study music perception. In this paper, we represent lyrics and chords in a shared vector space using a phrase-aligned lyrics-and-chords corpus and show that models that use these shared representations can predict musical genre of songs—a perceptual construct of music listening—better than models that do not use these representations. This work adds to our understanding of how lyrics and chords interact with one another in music and has applications in multimodal perception and music information retrieval.