- This event has passed.
AI Lecture Series – Mari Ostendorf
November 8 @ 2:00 pm - 3:00 pm
System Design Methodologies Professor
University of Washington
Friday, November 8, 2019
1:45pm – refreshments
2:00pm – lecture
Host: Ellen Riloff
Moving Beyond “Speech as Noisy Text” in Spoken Language Processing
Abstract: With substantial recent advances in automatic speech recognition (ASR) technology, there is growing interest in applying natural language processing (NLP) techniques to spoken language, with applications such as conversational agents and data mining of speech archives. Of course, despite the progress, ASR systems are not perfect, so NLP algorithms need to deal with noisy input. However, this pipelined view ignores a wealth of information in spoken language, which carries multiple sources of information (speaker identity, semantic meaning, speaker intent, emotion, etc.) and communicates it via multiple channels. Thus, it can be useful to treat the audio signal as multi-modal, where words and acoustic-prosodic cues (intonation, energy, timing) are different modalities. While prosody is well known to be important in speech synthesis, it has been a challenge to incorporate in speech understanding. However, recent experiments show that neural networks facilitate leveraging acoustic-prosodic cues, directly learning functions for processing F0 and energy. In addition, we find that accounting for the redundancy between syntax and prosody is important for multi-modal integration of words and acoustic features.
Bio: Mari Ostendorf joined the University of Washington in 1999. She is an Endowed Professor of System Design Methodologies in the Electrical & Computer Engineering Department, an Adjunct Professor in Linguistics and in Computer Science & Engineering, as well as Associate Vice Provost for Research. She is a Fellow of the IEEE, ISCA and ACL, a Scottish Informatics and Computer Science Alliance Distinguished Visiting Fellow, and a former Australian-American Fulbright Scholar. Prof. Ostendorf has published over 280 papers on a variety of topics in speech and language processing. In 2017, she served as a faculty advisor for the student team winning the inaugural Alexa Prize competition to build a socialbot, and conversational AI is a focus of her current work. Her research explores dynamic models for understanding and generating speech and text, particularly in multi-party contexts, and it contributes to a variety of applications, from education to clinical and scientific information extraction.