[2016-01-07] Dr. Yun-Nung (Vivian) Chen, Carnegie Mellon University, “Unsupervised Learning and Modeling of Knowledge and Intent for Spoken Dialogue Systems"

Poster:Post date:2015-12-30
Title: Unsupervised Learning and Modeling of Knowledge and Intent for Spoken Dialogue Systems
Date: 2016-01-07  1:00pm-2:00pm
Location: R210, CSIE
Speaker: Dr. Yun-Nung (Vivian) Chen, Carnegie Mellon University
Hosted by: Prof. Kun-Mao Chao


Various devices such as smartphones and in-car navigating systems are incorporating spoken dialogue systems (SDS) as personal intelligent assistants. Prior studies mostly focused on how to parse the natural language inputs into organized semantic concepts. Typically, such concepts and their structures are manually created for each application domain. Parsers that map utterances into these concepts are either manually crafted or automatically constructed from manually annotated corpora. Manual labor results in high cost, long duration and poor scalability of system development. To fill the gap, this talk focuses on improving generalization and scalability of building SDSs using general-purpose knowledge bases and by automatically inferring domain-specific semantic knowledge and structures from unlabeled conversations. We then propose a spoken language understanding (SLU) model for unifying the automatically acquired knowledge, decoding semantics, and predicting user intents simultaneously through a matrix factorization (MF) approach. Compared to other SLU models, our MF-SLU model achieves better and deeper understanding performance in terms of utterance semantics and user intents. In addition, the dissertation investigates the feasibility of applying the SLU modeling techniques developed for the human-machine dialogues to the human-human dialogues. The experiments show a great potential for reducing the cost of SDS development while achieving reasonable performance for both human-machine and human-human interactions.
Dr. Yun-Nung (Vivian) Chen 陳縕儂 received Ph.D. and M.S. degrees from Language Technologies Institute of School of Computer Science at Carnegie Mellon University, Pittsburgh, PA. Her research interests include spoken language understanding, spoken dialogue system, user modeling, speech summarization, information extraction, and machine learning. She received Best Student Paper Awards from IEEE ASRU 2013 and IEEE SLT 2010 and a Student Best Paper Nominee from INTERSPEECH 2012. Chen earned B.S. and M.S. degrees in Computer Science and Information Engineering from National Taiwan University, Taipei, Taiwan, where her master thesis received the ACLCLP Thesis Award. (http://vivianchen.idv.tw)

Last modification time:2015-12-30 PM 4:15

cron web_use_log