Automatic Speech Recognition & Learning

Automatic Speech Recognition & Learning  By Andrew Wickham

Speech recognition is a technology for language training that has been on the market for a number of years. It has been steadily improving with time and certain applications today are capable of relatively satisfactory evaluation of pronunciation and voice-to-text dictation. However, although these applications are able to understand simple sets of pre-recorded commands and recognise words and even simple phrases in some cases, they are as yet a long way from being able to interact with a speaker spontaneously or interpret meaning. For this reason, the term "speech recognition" is somewhat misleading, in my opinion.
 
In a typical speech recognition application,  learners listen to a question and are presented with three or more possible answers (in writing). The software is able in most cases to detect which answer the learner has read out. He is then presented with a model of the sentence he has pronounced (often a waveform), together with a correct model and expected to try to improve his pronunciation by imitating it.
 
In language learning, speech recognition applications may thus be useful for correcting and improving pronunciation (though there are still serious limits to this), but they are no substitute for real dialogue with a native speaker or a teacher. In the absence of a trainer, they do however oblige the learner to speak a lot and are a more motivating than the typical language laboratory courseware, since they provide at least the illusion of dialogue.
 

..........................................................

Andrew Wickham est un professionnel de la formation langues qui a fondé, puis dirigé de 1989 à 2003, la société Nexus Langues et Communication. Il a ensuite dirigé un projet de blended learning pour les cadres supérieurs de la société Renault. Il a à diverses occasions animé des débats de la profession. En 2008, il publie l'étude de marché sur la formation professionnelle linguistique en France (http://www.etude-langues.fr).

..........................................................

For more info see : 

Automatic Speech Recognition for second language learning: How and why it actually works
Ambra Neri, Catia Cucchiarini, and Wilhelmus Strik
A²RT, Department of Languge and Speech, University of Nijmegen, The Netherlands
{A.Neri, C.Cucchiarini, W.Strik}@let.kun.nl
SPEECH RECOGNITION FOR LANGUAGE TEACHING AND  EVALUATING: A STUDY OF EXISTING COMMERCIAL PRODUCTS
Rebecca Hincks, Centre for Speech Technology, Department of Speech, Music and Hearing
KTH, Stockholm, Sweden
hincks@speech.kth.se