The NTNU System at the Interspeech 2020 Non-Native Children's Speech ASR Challenge

TL;DR

This paper presents the NTNU ASR system for the challenging Interspeech 2020 Non-Native Children's Speech ASR Challenge, utilizing CNN-TDNNF models, data augmentation, and RNN language model rescoring to achieve competitive results.

Contribution

The paper introduces a robust ASR system for non-native children's speech using CNN-TDNNF models, advanced data augmentation, and language model rescoring under limited data conditions.

Findings

Achieved a 17.59% WER, ranking second in the challenge.

Effective use of data augmentation strategies improved recognition accuracy.

System outperformed the baseline and demonstrated robustness in challenging conditions.

Abstract

This paper describes the NTNU ASR system participating in the Interspeech 2020 Non-Native Children's Speech ASR Challenge supported by the SIG-CHILD group of ISCA. This ASR shared task is made much more challenging due to the coexisting diversity of non-native and children speaking characteristics. In the setting of closed-track evaluation, all participants were restricted to develop their systems merely based on the speech and text corpora provided by the organizer. To work around this under-resourced issue, we built our ASR system on top of CNN-TDNNF-based acoustic models, meanwhile harnessing the synergistic power of various data augmentation strategies, including both utterance- and word-level speed perturbation and spectrogram augmentation, alongside a simple yet effective data-cleansing approach. All variants of our ASR system employed an RNN-based language model to rescore the first-pass recognition hypotheses, which was trained solely on the text dataset released by the organizer. Our system with the best configuration came out in second place, resulting in a word error rate (WER) of 17.59 %, while those of the top-performing, second runner-up and official baseline systems are 15.67%, 18.71%, 35.09%, respectively.