TECHNIQUES FOR LEARNING EFFECTIVE MUSICAL FEATURES FOR GENERATIVE AND RETRIEVAL-BASED APPLICATIONS

Title:

TECHNIQUES FOR LEARNING EFFECTIVE MUSICAL FEATURES FOR GENERATIVE AND RETRIEVAL-BASED APPLICATIONS

Document Type and Number:

WIPO Patent Application WO/2021/029523

Kind Code:

A1

Abstract:

A method includes receiving a non-linguistic input associated with an input musical content. The method also includes, using a model that embeds multiple musical features describing different musical content and relationships between the different musical content in a latent space, identifying one or more embeddings based on the input musical content. The method further includes at least one of: (i) identifying stored musical content based on the one or more identified embeddings or (ii) generating derived musical content based on the one or more identified embeddings. In addition, the method includes presenting at least one of: the stored musical content or the derived musical content. The model is generated by training a machine learning system having one or more first neural network components and one or more second neural network components such that embeddings of the musical features in the latent space have a predefined distribution.

Inventors:

BRETAN PETER M (US)
HECK LARRY P (US)
JIN HONGXIA (US)

Application Number:

PCT/KR2020/007079

Publication Date:

February 18, 2021

Filing Date:

June 01, 2020

Export Citation:

Click for automatic bibliography generation Help

Assignee:

SAMSUNG ELECTRONICS CO LTD (KR)

International Classes:

G06Q50/10; G06F16/432; G06F16/65; G06N3/04; G06N3/08; G06V10/764

Domestic Patent References:

WO2019040132A1

2019-02-28

Foreign References:

US20180276540A1	2018-09-27
KR20180070340A	2018-06-26
KR20180066472A	2018-06-19
KR101934057B1	2019-04-08

Other References: