Senone-aware Adversarial Multi-task Training for Unsupervised Child to Adult Speech Adaptation

TL;DR

This paper introduces a senone-aware adversarial multi-task training method to improve unsupervised child speech adaptation, effectively reducing acoustic mismatch and enhancing performance across recognition and assessment tasks.

Contribution

It presents a novel adversarial multi-task training approach that leverages adult speech data to improve child speech modeling at the senone level.

Findings

7.7% relative error reduction in speech recognition

Up to 25.2% relative gains in child assessment tasks

Consistent outperformance of baseline methods

Abstract

Acoustic modeling for child speech is challenging due to the high acoustic variability caused by physiological differences in the vocal tract. The dearth of publicly available datasets makes the task more challenging. In this work, we propose a feature adaptation approach by exploiting adversarial multi-task training to minimize acoustic mismatch at the senone (tied triphone states) level between adult and child speech and leverage large amounts of transcribed adult speech. We validate the proposed method on three tasks: child speech recognition, child pronunciation assessment, and child fluency score prediction. Empirical results indicate that our proposed approach consistently outperforms competitive baselines, achieving 7.7% relative error reduction on speech recognition and up to 25.2% relative gains on the evaluation tasks.