Senone-aware Adversarial Multi-task Training for Unsupervised Child to Adult Speech Adaptation
This work addresses the challenge of child speech recognition and assessment, which is hindered by high acoustic variability and limited data, by leveraging adult speech resources, representing an incremental improvement over existing methods.
The paper tackled the problem of acoustic modeling for child speech by proposing a feature adaptation approach using adversarial multi-task training to reduce acoustic mismatch between adult and child speech at the senone level, achieving a 7.7% relative error reduction on speech recognition and up to 25.2% relative gains on assessment tasks.
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.