LANGUAGE IDENTIFICATION AND MULTILINGUAL SPEECH RECOGNITION USING DISCRIMINATIVELY TRAINED ACOUSTIC MODELS

Abstract We perform language identification experiments for four promi- nent South-African languages,using a multilingual speech recognition system. Specifically, we show how successfully Afrikaans, English, Xhosa and Zulu may be identified using a single set of HMMs and a single recognition pass. We further demonstrate,the effect of language,identification-specific dis- criminative acoustic model,training on both the per-language recognition accuracy as well as the accuracy of the language identification process. Experiments indicate that discriminative training leads to a small overall improvement in language iden- tification accuracy while not affecting the speech recognition performance strongly. Furthermore, language identification is found to be more error prone and discriminative training less effective for code-mixed utterances, indicating that these may require special treatment within a multilingual speech recogni- tion system.