Saved in:
Bibliographic Details
Main Authors: Wu, Junda, Li, Warren, Novack, Zachary, Namburi, Amit, Chen, Carol, McAuley, Julian
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.02271
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910630853738496
author Wu, Junda
Li, Warren
Novack, Zachary
Namburi, Amit
Chen, Carol
McAuley, Julian
author_facet Wu, Junda
Li, Warren
Novack, Zachary
Namburi, Amit
Chen, Carol
McAuley, Julian
contents Modeling temporal characteristics plays a significant role in the representation learning of audio waveform. We propose Contrastive Long-form Language-Audio Pretraining (\textbf{CoLLAP}) to significantly extend the perception window for both the input audio (up to 5 minutes) and the language descriptions (exceeding 250 words), while enabling contrastive learning across modalities and temporal dynamics. Leveraging recent Music-LLMs to generate long-form music captions for full-length songs, augmented with musical temporal structures, we collect 51.3K audio-text pairs derived from the large-scale AudioSet training dataset, where the average audio length reaches 288 seconds. We propose a novel contrastive learning architecture that fuses language representations with structured audio representations by segmenting each song into clips and extracting their embeddings. With an attention mechanism, we capture multimodal temporal correlations, allowing the model to automatically weigh and enhance the final fusion score for improved contrastive alignment. Finally, we develop two variants of the CoLLAP model with different types of backbone language models. Through comprehensive experiments on multiple long-form music-text retrieval datasets, we demonstrate consistent performance improvement in retrieval accuracy compared with baselines. We also show the pretrained CoLLAP models can be transferred to various music information retrieval tasks, with heterogeneous long-form multimodal contexts.
format Preprint
id arxiv_https___arxiv_org_abs_2410_02271
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle CoLLAP: Contrastive Long-form Language-Audio Pretraining with Musical Temporal Structure Augmentation
Wu, Junda
Li, Warren
Novack, Zachary
Namburi, Amit
Chen, Carol
McAuley, Julian
Sound
Artificial Intelligence
Audio and Speech Processing
Modeling temporal characteristics plays a significant role in the representation learning of audio waveform. We propose Contrastive Long-form Language-Audio Pretraining (\textbf{CoLLAP}) to significantly extend the perception window for both the input audio (up to 5 minutes) and the language descriptions (exceeding 250 words), while enabling contrastive learning across modalities and temporal dynamics. Leveraging recent Music-LLMs to generate long-form music captions for full-length songs, augmented with musical temporal structures, we collect 51.3K audio-text pairs derived from the large-scale AudioSet training dataset, where the average audio length reaches 288 seconds. We propose a novel contrastive learning architecture that fuses language representations with structured audio representations by segmenting each song into clips and extracting their embeddings. With an attention mechanism, we capture multimodal temporal correlations, allowing the model to automatically weigh and enhance the final fusion score for improved contrastive alignment. Finally, we develop two variants of the CoLLAP model with different types of backbone language models. Through comprehensive experiments on multiple long-form music-text retrieval datasets, we demonstrate consistent performance improvement in retrieval accuracy compared with baselines. We also show the pretrained CoLLAP models can be transferred to various music information retrieval tasks, with heterogeneous long-form multimodal contexts.
title CoLLAP: Contrastive Long-form Language-Audio Pretraining with Musical Temporal Structure Augmentation
topic Sound
Artificial Intelligence
Audio and Speech Processing
url https://arxiv.org/abs/2410.02271