To access this work you must either be on the Smith College campus OR have valid Smith login credentials.

On Campus users: To access this work if you are on campus please Select the Download button.

Off Campus users: To access this work from off campus, please select the Off-Campus button and enter your Smith username and password when prompted.

Non-Smith users: You may request this item through Interlibrary Loan at your own library.

Publication Date

2022-04-26

First Advisor

Katherine M. Kinnaird

Document Type

Honors Project

Degree Name

Bachelor of Arts

Department

Computer Science

Keywords

Music generation, Artificial intelligence, Deep learning, Music information retrieval

Abstract

In my thesis, I create and evaluate a deep learning autonomous music generation system with two layers. The first is a bidirectional Long Short Term Memory layer that carries out typical generation. The attention layer uses a novel approach to apply self-similarity matrices to previous time steps. This attention mechanism imposes an input structure on the generated music. I train this model on the MAESTRO dataset, and compare its performance on new data to the same model without the attention mechanism. The addition of the attention mechanism significantly improves the network’s ability to replicate specific structures, and it performs comparatively better on a validation dataset than a model without the attention mechanism.

Rights

©2022 Sophia Hager. Access limited to the Smith College community and other researchers while on campus. Smith College community members also may access from off-campus using a Smith College log-in. Other off-campus researchers may request a copy through Interlibrary Loan for personal use.

Language

English

Download

Smith Only:
Off Campus Download

Share

COinS