VoxPopuli: A Large-Scale Multilingual Speech Corpus for Representation Learning, Semi-Supervised Learning and Interpretation

TL;DR

VoxPopuli is a comprehensive multilingual speech corpus with 100K hours of unlabelled data across 23 languages, enabling advances in unsupervised and semi-supervised speech representation learning and interpretation.

Contribution

It introduces the largest open multilingual speech dataset with extensive unlabelled and transcribed data, facilitating research in unsupervised and semi-supervised speech processing.

Findings

Validated semi-supervised learning benefits with VoxPopuli data

Provided speech recognition baselines for multiple languages

Demonstrated effectiveness in out-of-domain settings

Abstract

We introduce VoxPopuli, a large-scale multilingual corpus providing 100K hours of unlabelled speech data in 23 languages. It is the largest open data to date for unsupervised representation learning as well as semi-supervised learning. VoxPopuli also contains 1.8K hours of transcribed speeches in 16 languages and their aligned oral interpretations into 5 other languages totaling 5.1K hours. We provide speech recognition baselines and validate the versatility of VoxPopuli unlabelled data in semi-supervised learning under challenging out-of-domain settings. We will release the corpus at https://github.com/facebookresearch/voxpopuli under an open license.