Understanding systems: a survey
Frédéric Béchet
University of Avignon, LIA
France
In commercial applications based on speech recognition, the quality of the human-computer interaction is still far from being effective both from the user and service provider side. To improve the effectiveness and user acceptance of automated dialogue systems it is necessary to advance the state of the art of Spoken Language Understanding (SLU) research in two directions. Firstly, SLU models have to be tightly coupled with the upstream (Automatic Speech Recognition: ASR) and downstream (Dialog Management: DM) processes. Secondly, SLU models have to be part of an adaptive component whose parameters are updated on-line based on the outcome of dialog strategies, a-priori or a-posteriori knowledge.
This talk will present a survey on SLU systems following these two directions. The focus of this talk will be on the robust processing of spontaneous speech in "real life" applications that have to deal with a large variety of speakers with different user experiences ("novice" vs. "expert" users). The plan of this survey will follow a set of questions that can help define this domain at the junction between speech processing on one side and natural language processing on the other side:
- What kind of formal representation can be used in order to represent the meaning of an utterance?
- Which models and which algorithms can be used to develop a robust SLU system?
- On what kind of corpus and with which annotations can we train and develop a SLU system?
- How can we evaluate the performance of a SLU system?
Frederic Bechet is a researcher in the field of Speech and Natural Language Processing at the computer laboratory (LIA) of the University of Avignon, France. He was an invited professor for one year at the AT&T Research Lab in Florham Park, New Jersey, USA, working within the "How May I Help You?" research project. Frederic Bechet is the author/co-author of over 50 refereed papers for journals and international conferences. He is currently a member (2008-2010) of the IEEE Signal Processing Society Speech and Language Technical Committee (SLTC). His research activities are mainly focused on Spoken Language Understanding for both Spoken Dialogue Systems and Speech Mining applications.