Data Augmentation for Spoken Grammatical Error Correction

TL;DR

This paper introduces an automated method to generate augmented audio-text datasets with grammatical errors and disfluencies for spoken GEC, aiming to improve data resources for low-resource spoken language correction tasks.

Contribution

It proposes a novel automated data augmentation technique for spoken GEC and introduces objective metrics to evaluate and select suitable augmented datasets.

Findings

Augmented datasets improve GEC performance on speech data.

The method maintains original data characteristics while adding diverse errors.

Evaluation metrics effectively identify high-quality augmented data.

Abstract

While there exist strong benchmark datasets for grammatical error correction (GEC), high-quality annotated spoken datasets for Spoken GEC (SGEC) are still under-resourced. In this paper, we propose a fully automated method to generate audio-text pairs with grammatical errors and disfluencies. Moreover, we propose a series of objective metrics that can be used to evaluate the generated data and choose the more suitable dataset for SGEC. The goal is to generate an augmented dataset that maintains the textual and acoustic characteristics of the original data while providing new types of errors. This augmented dataset should augment and enrich the original corpus without altering the language assessment scores of the second language (L2) learners. We evaluate the use of the augmented corpus both for written GEC (the text part) and for SGEC (the audio-text pairs). Our experiments are conducted on the S\&I Corpus, the first publicly available speech dataset with grammar error annotations.