Home // International Journal On Advances in Internet Technology, volume 8, numbers 1 and 2, 2015 // View article

Services to Support Use and Development of Multilingual Speech Input

Authors:
António Teixeira
Pedro Francisco
Nuno Almeida
Carlos Pereira
Samuel Silva

Keywords: Services; speech; semantic grammars; multilingual; multimodal interaction; speech recognition; translation

Abstract:
The use of speech for human-machine interaction benefits from being our most natural way of communication. Profiting from a widespread use of devices providing audio input, applications increasingly support speech recognition and understanding, but several challenges are still to be addressed. A common solution to support spoken language understanding uses semantic grammars. Also, for speech recognizers, grammars are a common way of configuring what can be recognized. The definition of these grammars, to support a semantically rich spoken interaction, involves a large amount of resources and linguistic knowledge and, therefore, systems tend to support a single language or multiple languages in a very narrow semantic scope. To address this issue, we propose to use an existing grammar to generate grammars for other languages by using automatic translation and taking into consideration aspects such as word realignment and grammar expansion, based on multiple possible translations. This method fosters effortless creation of a semantically rich grammar, for the target language, which can then be revised. The proposed method has already been successfully used to enable multilingual speech interaction for AALFred, a personal life assistant, covering English, French, Hungarian, Polish, and Portuguese.

Pages: 1 to 12

Copyright: Copyright (c) to authors, 2015. Used with permission.

Publication date: June 30, 2015

Published in: journal

ISSN: 1942-2652