End-to-end neural systems for automatic children speech recognition: An empirical study. (March 2022)

• Record Type: Journal Article
• Authors: Gurunath Shivakumar, Prashanth; Narayanan, Shrikanth
• Is Part Of: Computer speech & language. Volume 72(2022)
• Publication Date: 2022-03
• Subjects: Children speech recognition -- End-to-end speech recognition -- Residual network -- Time depth separable convolutional network -- Transformer; Speech processing systems -- Periodicals; Automatic speech recognition -- Periodicals; Computers -- Periodicals; Linguistics -- Periodicals; Speech-Language Pathology -- Periodicals; Traitement automatique de la parole -- Périodiques; Reconnaissance automatique de la parole -- Périodiques; Automatic speech recognition; Speech processing systems; Electronic journals; Periodicals; 006.454
• DOI: 10.1016/j.csl.2021.101289
• Languages: English
• ISSNs: 0885-2308

Abstract

Abstract: A key desiderata for inclusive and accessible speech recognition technology is ensuring its robust performance to children's speech. Notably, this includes the rapidly advancing neural network based end-to-end speech recognition systems. Children speech recognition is more challenging due to the larger intra-inter speaker variability in terms of acoustic and linguistic characteristics compared to adult speech. Furthermore, the lack of adequate and appropriate children speech resources adds to the challenge of designing robust end-to-end neural architectures. This study provides a critical assessment of automatic children speech recognition through an empirical study of contemporary state-of-the-art end-to-end speech recognition systems. Insights are provided on the aspects of training data requirements, adaptation on children data, and the effect of children age, utterance lengths, different architectures and loss functions for end-to-end systems and role of language models on the speech recognition performance. Highlights: Empirical study of end-to-end deep learning based children speech ASR Training/Adaptation comparisons for various adult and children speech dataset sizes Acoustic Model Comparison: DNN-HMM hybrid vs. End-to-End ASR DNN Architecture Comparison: ResNets vs. Time-depth separable CNN vs. Transformers Loss Function Comparison: Connectionist Temporal Classification vs. Seq-to-Seq Language Model Comparison: 4gram vs. 6gram vs. Gated-CNN LM; word vs. word-piece LM Error Analysis: Letter Errors vs. Word Errors; Utterance Lengths; Confusion Analysis