Séminaire de M. Basem Ahmad, doctorant de l'Université SAINS Malaisie - Date : jeudi 17 mai 2012, 14h00 - Lieu : "seminar room", 9ème étage, B1, Institut MICA, Hanoi University of Science and Technology

Intervenant :
M. Basem Ahmad, doctorant de l'Universite SAINS, Penang, Malaisie

Date : jeudi 17 mai 2012, 14h00
Lieu : salle "seminar room", 9ème étage, Institut MICA
Interprète traducteur : le séminiare sera présenté en anglais

Résumé/astract:
Speech is the most convenient and direct way people communicate. There are more than six thousand languages reported in the world today. Many people nowadays know two or more languages. The present work proposes an automatic speech recognition (ASR) task for code-switching speech. Code-switching speech involves using more than one language within an utterance or discourse. There are many reasons why a speaker code switch when they speak. Individuals who live in a society where more than one language is spoken usually code-switch occur. For example, when a Malay speaker speaks Malay, switching to English words or phrases is easier than thinking of the correct words in Malay. According to reports, ethnic minority communities can retain their cultural identity with the help of code switching. Often, people also speak in another language to establish friendly relationships with others whose mother tongue is another language. The code-switching speaking style can usually be found in many bilingual or multilingual societies. Examples of these societies are the French–German in Switzerland, the English–Spanish in the US, the Cantonese–English in Hong Kong, and the Mandarin–Taiwanese in Taiwan, the Malay–English and Mandarin-English in Malaysia. This presentation will focus on our research in pronunciation modeling and acoustic modeling to improve the code-switching ASR.