Séminaire de Dr Thang Vu, Institute for Natural Language Processing (IMS), University of Stuttgart - Date : jeudi 18 février 2016, 14h00 - Lieu : salle "seminar", Institut MICA, Hanoi University of Science and Technology

Intervenant :
Dr Thang VU, researcher of the Institute for Natural Language Processing (IMS), University of Stuttgart

Date : jeudi 18 février 2016, 14h00
Lieu : salle "seminar room", 9ème étage, bâtiment B1, Institut MICA, Hanoi University of Science and Technology
Langue : le séminaire sera présenté en anglais

Résumé/abstract:
The performance of speech and language processing technologies has improved significantly in the last few years with the use of deep learning. It opens a new exciting chapter of speech and language related research. Multilingual speech recognition and language understanding are among the most important open research questions nowadays. Although the fact is known that there are more than 7000 languages in the world, most efforts to date, however, were focused only on a very small number of languages. Hence, supporting multiple input and output languages is a real need. Furthermore with the use of deep learning, speech recognition is getting closer to a state where it can be used more widely in the daily life. Therefore, language understanding is one of the next essential steps. In this talk, I will present techniques for rapidly building speech recognition systems for low-resource languages using multilingual and crosslingual information. This includes the sharing of data and models across languages based on deep learning. Furthermore, I will report ongoing research towards spoken language understanding (e.g. slot filling and sentiment analysis) at the Institute for Natural Language Processing (IMS), University Stuttgart.