GraphMuse: A Library for Symbolic Music Graph Processing

TL;DR

GraphMuse is a new library that streamlines symbolic music graph processing and GNN training, introducing a music-specific neighbor sampling method and hierarchical modeling to improve musical task performance.

Contribution

It introduces a unified framework with novel neighbor sampling and hierarchical modeling tailored for symbolic music graph analysis.

Findings

Significant performance improvements in pitch spelling and cadence detection tasks.

Enhanced expressivity of graph networks for musical applications.

Facilitates standardization and progress in symbolic music graph processing.

Abstract

Graph Neural Networks (GNNs) have recently gained traction in symbolic music tasks, yet a lack of a unified framework impedes progress. Addressing this gap, we present GraphMuse, a graph processing framework and library that facilitates efficient music graph processing and GNN training for symbolic music tasks. Central to our contribution is a new neighbor sampling technique specifically targeted toward meaningful behavior in musical scores. Additionally, GraphMuse integrates hierarchical modeling elements that augment the expressivity and capabilities of graph networks for musical tasks. Experiments with two specific musical prediction tasks -- pitch spelling and cadence detection -- demonstrate significant performance improvement over previous methods. Our hope is that GraphMuse will lead to a boost in, and standardization of, symbolic music processing based on graph representations. The library is available at https://github.com/manoskary/graphmuse