Unrolled Creative Adversarial Network For Generating Novel Musical Pieces

TL;DR

This paper introduces unrolled Creative Adversarial Networks (CAN) for music generation, addressing mode collapse and enabling the creation of both style-agnostic and style-deviating novel musical pieces.

Contribution

It extends the CAN framework to music, proposing unrolled CAN to improve diversity and creativity in generated compositions.

Findings

Unrolled CAN reduces mode collapse in music generation.

The system can generate both style-neutral and style-deviating music.

Evaluation shows increased creativity and variation in outputs.

Abstract

Music generation has emerged as a significant topic in artificial intelligence and machine learning. While recurrent neural networks (RNNs) have been widely employed for sequence generation, generative adversarial networks (GANs) remain relatively underexplored in this domain. This paper presents two systems based on adversarial networks for music generation. The first system learns a set of music pieces without differentiating between styles, while the second system focuses on learning and deviating from specific composers' styles to create innovative music. By extending the Creative Adversarial Networks (CAN) framework to the music domain, this work introduces unrolled CAN to address mode collapse, evaluating both GAN and CAN in terms of creativity and variation.