LyricLens: An Interactive System for Multi-Label Music Content Rating

Kai-Yu Lu, Malhar Sham Ghogare, Zihan Su, Shanu Sushmita

Keywords: Natural Language Processing, Multi-label Classification, Music Content Rating (MCR), Song Lyrics Analysis

Abstract:

This paper presents LyricLens, the first online, multi-label lyric classification and rating system. The system's core is a novel classifier, trained on a large-scale Spotify dataset, that identifies four categories of explicit content: Sexual Content, Violence, Language, and Substance Use.
LyricLens introduces the Music Content Rating (MCR) framework, a nuanced, five-level certification system (M-E, M-P, M-T, M-R, M-AO) that replaces traditional binary explicit tags. This framework, which also provides specific content descriptors, adapts established rating standards from movies and video games for use with lyrical content. LyricLens is a valuable tool for parental oversight, automated content moderation, and academic research. By providing a detailed, granular analysis of lyrics, the system promotes a more responsible and informed consumption of music, especially for young audiences.

Track: Demo Track

Confirmation: Demo Track: I confirm that I have followed the formatting guideline and included all author names and affiliations.

(Optional) Short Video Recording Link: https://www.youtube.com/watch?v=nmLQigBSWjE

Submission Number: 46