MUSYNERGY: A framework for music collaboration discovery based on neural networks and graph analysis

Abstract

The music industry has been reshaped by the rise of artist collaborations, driven by digital technologies, streaming platforms, and the globalization of music. While existing research has examined the cultural and commercial impact of collaborations, few efforts have focused on recommendation systems to assist musicians in discovering potential creative partners. Moreover, most approaches rely on proprietary data, limiting scalability and reproducibility. This paper presents MUSYNERGY, a novel framework for music collaboration discovery based on neural networks and graph analysis. MUSYNERGY builds a Heterogeneous Knowledge Graph (HKG) using open data from MusicBrainz, representing relationships among artists, tracks, and musical attributes over five decades. By formulating collaboration discovery as a link prediction task, the system identifies new, plausible collaborations between artists with no prior joint work. This open, scalable framework addresses current limitations in data accessibility and supports innovation, transparency, and cultural exchange in the global music landscape through data-driven collaboration discovery.