A customized residual neural network and bi-directional gated recurrent unit-based automatic speech recognition model

• Tipo: Artigo em Revista Científica Internacional
• Data: 2022-04

Autores

Selim Reza (Autor) Outra; Marta Campos Ferreira (Autor) FEUP; J.J.M. Machado (Autor) FEUP; João Manuel R. S. Tavares (Autor) FEUP

Revista

Título: Expert Systems with Applications

Vol. 215 Nº 119293

Páginas: 1-10

ISSN: 0957-4174

Editora: Elsevier

Indexação

ISI Web of Knowledge - 12 Citações

ISI Web of Science

Scopus - 19 Citações

Clarivate Analytics

Classificação Científica

CORDIS: Ciências Tecnológicas

FOS: Ciências da engenharia e tecnologias

Projetos Associados

Safe Cities - Safe Cities - Inovação para Construir Cidades Seguras

Outras Informações

ID Authenticus: P-00X-FZ7

DOI: 10.1016/j.eswa.2022.119293

Abstract (EN): Speech recognition aims to convert human speech into text and has applications in security, healthcare, commerce, automobiles, and technology, just to name a few. Inserting residual neural networks before recurrent neural network cells improves accuracy and cuts training time by a good margin. Furthermore, layer normalization instead of batch normalization is more effective in model training and performance enhancement. Also, the size of the datasets presents tremendous influences in achieving the best performance. Leveraging these tricks, this article proposes an automatic speech recognition model with a stacked five layers of customized Residual Convolution Neural Network and seven layers of Bi-Directional Gated Recurrent Units, including a logarithmic so f tmax for the model output. Each of them incorporates a learnable per-element affine parameter-based layer normalization technique. The training and testing of the new model were conducted on the LibriSpeech corpus and LJ Speech dataset. The experimental results demonstrate a character error rate (CER) of 4.7 and 3.61% on the two datasets, respectively, with only 33 million parameters without the requirement of any external language model.

Idioma: Inglês

Tipo (Avaliação Docente): Científica

Nº de páginas: 10