Multimodal and multi-label classification of music genre

• Grant number: 20/07911-2
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: August 01, 2020
• End date: July 31, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Diego Furtado Silva
• Grantee: Luís Felipe Corrêa Ortolan
• Host Institution: Centro de Ciências Exatas e de Tecnologia (CCET). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil

Abstract

With the rapid growth in the distribution of digital music, organizing and retrieving musical information through computational methods has become an increasingly relevant task. While the volume of data associated with music collections becomes larger and more diverse, providing a good experience for users becomes a more complex task. The classification of musical genres is an important task for improving the user experience with music. However, subjectivity in the conceptualization of musical genre makes its definition quite a difficult task. While most works in this application are limited to the content of the audio, there are works that show that the lyrics of a song can contain information relevant to the classification of musical genre. These factors suggest that it is possible to use the lyrics in conjunction with its content to perform the multimodal classification of musical genre. Thus, this project will investigate the fusion of machine learning techniques and natural language processing for the classification of the musical genre of artists and songs from audio data and associated lyrics. (AU)