ASAICLLGMMJul 27, 2023

Emotion4MIDI: a Lyrics-based Emotion-Labeled Symbolic Music Dataset

arXiv:2307.14783v12 citationsh-index: 5
Originality Incremental advance
AI Analysis

This provides a valuable resource for researchers in music generation and emotion analysis to develop models that generate music based on specific emotions.

The researchers tackled the lack of emotion-labeled symbolic music data by creating a large-scale dataset of 12k MIDI songs with fine-grained emotion labels derived from lyrics, using a state-of-the-art emotion classification model that is half the size of the baseline.

We present a new large-scale emotion-labeled symbolic music dataset consisting of 12k MIDI songs. To create this dataset, we first trained emotion classification models on the GoEmotions dataset, achieving state-of-the-art results with a model half the size of the baseline. We then applied these models to lyrics from two large-scale MIDI datasets. Our dataset covers a wide range of fine-grained emotions, providing a valuable resource to explore the connection between music and emotions and, especially, to develop models that can generate music based on specific emotions. Our code for inference, trained models, and datasets are available online.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes