The Arts

Exploring how a Generative AI interprets music

G. Barenboim, L. D. Debbio, et al.

Discover how Google's MusicVAE interprets music, revealing fascinating insights into 'music neurons' and how they distinguish elements like pitch, rhythm, and melody. This groundbreaking research was conducted by Gabriela Barenboim, Luigi Del Debbio, Johannes Hirn, and Verónica Sanz.

00:00

Playback language: English

Index

Abstract

This paper investigates how Google's MusicVAE, a Variational Auto-Encoder, interprets music. The researchers analyze the latent space of the model, identifying 'music neurons' that encode musical information and 'noise neurons' that remain largely inactive. They explore how pitch, rhythm, and melody are represented within the music neurons, finding that pitch and rhythm are primarily encoded in the first few neurons, while melody becomes more apparent in independent neurons for longer musical sequences.

Publisher

Springer Nature

Published On

Jan 01, 2023

Authors

Gabriela Barenboim, Luigi Del Debbio, Johannes Hirn, Verónica Sanz

Related Publications

Explore these studies to deepen your understanding of the subject.

Business

Generative AI as a catalyst for HRM practices: mediating effects of trust

K. D. V. Prasad and T. De

Computer Science

Exploring the mechanism of sustained consumer trust in AI chatbots after service failures: a perspective based on attribution and CASA theories

C. Gu, Y. Zhang, et al.

Computer Science

Exploring excitement counterbalanced by concerns towards AI technology using a descriptive-prescriptive data processing method

S. Oprea and A. Bâra

Chemistry

Diffusion-based generative AI for exploring transition states from 2D molecular graphs

S. Kim, J. Woo, et al.

Listen, Learn & Level Up

Over 10,000 hours of research content in 25+ fields, available in 12+ languages.

No more digging through PDFs, just hit play and absorb the world's latest research in your language, on your time.

listen to research audio papers with researchbunny