Review-Driven Multi-Label Music Style Classification by Exploiting Style Correlations

TL;DR

This paper introduces a deep learning approach for review-driven multi-label music style classification, effectively capturing style correlations to improve accuracy over existing methods.

Contribution

It proposes a novel label-graph neural network and correlation-based training mechanism specifically designed for multi-label music style classification from reviews.

Findings

Micro F1 score improved from 53.9 to 64.5

One-error reduced from 30.5 to 22.6

Approach effectively captures style correlations

Abstract

This paper explores a new natural language processing task, review-driven multi-label music style classification. This task requires the system to identify multiple styles of music based on its reviews on websites. The biggest challenge lies in the complicated relations of music styles. It has brought failure to many multi-label classification methods. To tackle this problem, we propose a novel deep learning approach to automatically learn and exploit style correlations. The proposed method consists of two parts: a label-graph based neural network, and a soft training mechanism with correlation-based continuous label representation. Experimental results show that our approach achieves large improvements over the baselines on the proposed dataset. Especially, the micro F1 is improved from 53.9 to 64.5, and the one-error is reduced from 30.5 to 22.6. Furthermore, the visualized analysis shows that our approach performs well in capturing style correlations.