Mispronunciation Detection Using Neural Networks for Second Language Learners

Abstract

Speaking a second language fluently is the aim of any language learner. Computer-aided language learning (CALL) systems help learners achieve this goal. Mispronunciation detection can be considered the most helpful component in CALL systems. For this reason, the focus is currently on research in mispronunciation detection systems. There are different methods for mispronunciation detection, such as posterior probability-based methods and classifier-based methods. Recently, deep-learning-based methods have also attracted great interest and are being studied. This chapter reviews the research that proposed neural network methods for mispronunciation detection conducted between 2014 and 2021 for second language learners. The results obtained from studies in the literature and comparisons between different techniques are also discussed.

Key Terms in this Chapter

Mispronunciation: The incorrect pronunciation of a specific word or sound.

Neural Network: A series of algorithms that aims to recognize underlying relationships in a set of data through a process that mimics how the human brain operates.

Corpus: A language dataset selected systematically and stored as an electronic database.

Deep Neural Network: A neural network with more than two layers.

Goodness of Pronunciation: A method used for automatic mispronunciation detection.

Corpora: Plural of the corpus.

Computer-Assisted Language Learning: An approach to teach and learning within which the pc and computer-based resources like the web are used to assess the material to be discovered. It includes considerable interactive elements.