WavCaps: A ChatGPT-Assisted Weakly-Labelled Audio Captioning Dataset for Audio-Language Multimodal Research

University of Surrey · Johns Hopkins University · +2 more institutions

Indexed incrossref

Abstract

The advancement of audio-language (AL) multimodal learning tasks has been significant in recent years, yet the limited size of existing audio-language datasets poses challenges for researchers due to the costly and time-consuming collection process. To address this data scarcity issue, we introduce WavCaps, the first large-scale weakly-labelled audio captioning dataset, comprising approximately 400 k audio clips with paired captions. We sourced audio clips and their raw descriptions from web sources and a sound event detection dataset. However, the online-harvested raw descriptions are highly noisy and unsuitable for direct use in tasks such as automated audio captioning. To overcome this issue, we propose a…

Citation impact

123
total citations
FWCI
38.88
Percentile
100%
References
106
Citations per year

Authors

9

Topics & keywords

Keywords
  • Closed captioning
  • Computer science
  • Speech recognition
  • Audio analyzer
  • Natural language processing
  • Linguistics
  • Artificial intelligence
  • Audio signal
No related works found for this paper.

Funding