Lyrics Matter: Exploiting the Power of Learnt Representations for Music Popularity Prediction

TL;DR

This paper introduces a multimodal approach combining lyrics, audio, and social data using large language models to improve music popularity prediction, significantly outperforming existing methods on a large dataset.

Contribution

It presents a novel pipeline for extracting high-dimensional lyric embeddings with LLMs and integrates them into a multimodal model for better popularity prediction.

Findings

Achieves 9% reduction in MAE and 20% in MSE over baselines.

Demonstrates the effectiveness of LLM-driven lyric features.

Validates the importance of dense lyric representations through ablation studies.

Abstract

Accurately predicting music popularity is a critical challenge in the music industry, offering benefits to artists, producers, and streaming platforms. Prior research has largely focused on audio features, social metadata, or model architectures. This work addresses the under-explored role of lyrics in predicting popularity. We present an automated pipeline that uses LLM to extract high-dimensional lyric embeddings, capturing semantic, syntactic, and sequential information. These features are integrated into HitMusicLyricNet, a multimodal architecture that combines audio, lyrics, and social metadata for popularity score prediction in the range 0-100. Our method outperforms existing baselines on the SpotGenTrack dataset, which contains over 100,000 tracks, achieving 9% and 20% improvements in MAE and MSE, respectively. Ablation confirms that gains arise from our LLM-driven lyrics feature pipeline (LyricsAENet), underscoring the value of dense lyric representations.