Libri-Adapt: A New Speech Dataset for Unsupervised Domain Adaptation

TL;DR

Libri-Adapt is a new speech dataset derived from LibriSpeech, designed to evaluate and improve unsupervised domain adaptation in speech recognition across diverse real-world acoustic and hardware conditions.

Contribution

The paper introduces Libri-Adapt, a comprehensive dataset for studying domain shifts in ASR, and provides baseline results demonstrating the impact of various domain variations.

Findings

Domain shifts significantly affect ASR performance.

Libri-Adapt enables systematic evaluation of domain adaptation methods.

Baseline results highlight challenges in real-world speech recognition.

Abstract

This paper introduces a new dataset, Libri-Adapt, to support unsupervised domain adaptation research on speech recognition models. Built on top of the LibriSpeech corpus, Libri-Adapt contains English speech recorded on mobile and embedded-scale microphones, and spans 72 different domains that are representative of the challenging practical scenarios encountered by ASR models. More specifically, Libri-Adapt facilitates the study of domain shifts in ASR models caused by a) different acoustic environments, b) variations in speaker accents, c) heterogeneity in the hardware and platform software of the microphones, and d) a combination of the aforementioned three shifts. We also provide a number of baseline results quantifying the impact of these domain shifts on the Mozilla DeepSpeech2 ASR model.