Language Identification for Under-ResourcedLanguages in the Basque Context

摘要

Automatic Speech Recognition (ASR) is a broad research area that ab-sorbs many efforts from the research community. The interest on MultilingualSystems arouses in the Basque Country because there are three official languages(Basque, Spanish, and French), and there is much linguistic interaction amongthem, even if Basque has very different roots than the other two languages. Thedevelopment of Multilingual Large Vocabulary Continuous Speech Recognitionsystems involves issues as: Language Identification, Acoustic Phonetic Decoding,Language Modeling or the development of appropriate Language Resources. Thispaper describes the development of a Language Identification (LID) system ori-ented to robust Multilingual Speech Recognition in the Basque context. The workpresents hybrid strategies for LID, based on the selection of system elements byseveral classifiers and Discriminant Analysis improved with robust regularizedcovariance matrix estimation methods oriented to under-resourced languages andstochastic methods for speech recognition tasks (Hidden Markov Modelsand n-grams).