Estimating Musical Surprisal from Audio in Autoregressive Diffusion Model Noise Spaces

TL;DR

This paper explores the use of autoregressive diffusion models to estimate musical surprisal from audio, demonstrating their effectiveness in modeling musical expectations and outperforming previous transformer-based methods in key tasks.

Contribution

It introduces diffusion model-based information content estimation for musical surprisal, showing improved performance over transformer models in audio prediction tasks.

Findings

Diffusion models better describe data than GIVT in negative log-likelihood.

Diffusion IC effectively captures musical surprisal and segment boundaries.

Surprisal estimates vary with noise levels, aligning with audio feature granularities.

Abstract

Recently, the information content (IC) of predictions from a Generative Infinite-Vocabulary Transformer (GIVT) has been used to model musical expectancy and surprisal in audio. We investigate the effectiveness of such modelling using IC calculated with autoregressive diffusion models (ADMs). We empirically show that IC estimates of models based on two different diffusion ordinary differential equations (ODEs) describe diverse data better, in terms of negative log-likelihood, than a GIVT. We evaluate diffusion model IC's effectiveness in capturing surprisal aspects by examining two tasks: (1) capturing monophonic pitch surprisal, and (2) detecting segment boundaries in multi-track audio. In both tasks, the diffusion models match or exceed the performance of a GIVT. We hypothesize that the surprisal estimated at different diffusion process noise levels corresponds to the surprisal of music and audio features present at different audio granularities. Testing our hypothesis, we find that, for appropriate noise levels, the studied musical surprisal tasks' results improve. Code is provided on github.com/SonyCSLParis/audioic.