Abstract

Interacting with machines via natural language has been an emerging trend. Recent advances in deep learning enabled new research frontiers for goal-oriented neural dialogue systems. This talk will cover recent research about neural dialogue systems, with components of natural language understanding (NLU), dialogue policy, and natural language generation (NLG). In NLU, a slot gate is proposed to learn the relationship between the intent and slots to obtain better semantic frame results by the global optimization. The history utterances and their temporal information are leveraged for better contextual understanding. In policy learning, we propose D3Q (Discriminative Deep Dyna-Qj, which integrates planning to efficiently utilize real-user experiences and incorporates a discriminator to improve the learning robustness. In NLG, we introduce a hierarchical decoding model based on linguistic patterns in different levels to tackle the issues about generating shorter and grammatically incorrect sentences.

Speaker Biography

Yun-Nung (Vivian) Chen is currently an assistant professor at the Department of Computer Science & Information Engineering, National Taiwan University. She earned her Ph.D. degree from Carnegie Mellon University, where her research interests focus on spoken dialogue system, language understanding, natural language processing, and multi-modal speech application. She received MOST Young Scholar Fellowship 2018, Google Faculty Award 2016, two Student Best Paper Awards from IEEE SLT 2010 and IEEE ASRU 2013, a Student Best Paper Nominee from Interspeech 2012, and the Distinguished Master Thesis Award from ACLCLP. Prior to joining National Taiwan University, she worked in the Deep Learning Technology Center at Microsoft Research Redmond. More information about her can be found at http://vivianchen.idv.tw.