Unsupervised Active Learning: Optimizing Labeling Cost-Effectiveness for Automatic Speech Recognition

TL;DR

This paper introduces an unsupervised active learning framework for automatic speech recognition that reduces labeling costs by selecting the most informative data, achieving significant WER improvements or cost savings.

Contribution

It proposes a novel unsupervised data selection method using contrastive techniques to optimize labeling efficiency in speech recognition tasks.

Findings

Improves WER by over 11% with the same labeled data

Halves labeling costs while maintaining WER

Effective data selection framework for ASR

Abstract

In recent years, speech-based self-supervised learning (SSL) has made significant progress in various tasks, including automatic speech recognition (ASR). An ASR model with decent performance can be realized by fine-tuning an SSL model with a small fraction of labeled data. Reducing the demand for labeled data is always of great practical value. In this paper, we further extend the use of SSL to cut down labeling costs with active learning. Three types of units on different granularities are derived from speech signals in an unsupervised way, and their effects are compared by applying a contrastive data selection method. The experimental results show that our proposed data selection framework can effectively improve the word error rate (WER) by more than 11% with the same amount of labeled data, or halve the labeling cost while maintaining the same WER, compared to random selection.