8 Papers You Can’t Miss From ISMIR2021 (AI Music Conference)

In November, I had the chance to attend ISMIR2021 (International Society of Music Information Retrieval Conference), the most important conference in AI Music. If you’re into AI Music, you should always keep an eye on this conference. This year’s iteration was particularly interesting, because of the breadth of the topics covered, from music cover detection, to music representations, from generative music, to music psychology experiments. In this video, I share an overview of the 8 papers I liked the most. 0:00 Intro 2:10 The Words Remain the Same: Cover Detection with Lyrics Transcription 8:05 Cross-cultural Mood Perception in Pop Songs and its Alignment with Mood Detection Algorithm 11:34 A Modular System for the Harmonic Analysis of Musical Scores using a Large Vocabulary 19:30 Does Track Sequence in User-generated Playlists Matter? 25:40 Semi-supervised Music Tagging Transformer 34:26 Controllable Deep Melody Generation via Hierarchical Music Structure Rep