An Autoethnographic Exploration of XAI in Algorithmic Composition

TL;DR

This paper presents an autoethnographic study of using an explainable AI model for music generation, revealing how interpretability influences musical feature exploration and workflow integration.

Contribution

It introduces the first practical exploration of a generative XAI model in music making, demonstrating its role in enhancing creative workflows.

Findings

XAI models highlight dataset features over model features

XAI integration enriches iterative music-making workflows

Interpretability aids in understanding generative music processes

Abstract

Machine Learning models are capable of generating complex music across a range of genres from folk to classical music. However, current generative music AI models are typically difficult to understand and control in meaningful ways. Whilst research has started to explore how explainable AI (XAI) generative models might be created for music, no generative XAI models have been studied in music making practice. This paper introduces an autoethnographic study of the use of the MeasureVAE generative music XAI model with interpretable latent dimensions trained on Irish folk music. Findings suggest that the exploratory nature of the music-making workflow foregrounds musical features of the training dataset rather than features of the generative model itself. The appropriation of an XAI model within an iterative workflow highlights the potential of XAI models to form part of a richer and more complex workflow than they were initially designed for.