Event Details

Conversational AI is currently at a tipping point. On the one hand, there is growing consensus on the promise of conversational AI as a universal interface that seamlessly integrates a wide range of backend data and services and builds trust with users via informative and engaging interaction. On the other hand, research on conversational AI has yet met the high hopes, and mainstream conversational AI agents still present significant limitations in expressiveness, contextuality, and trustworthiness, among others.

In this talk, I will discuss several emerging frontiers of conversational AI research and applications:

1. Representation. I will describe a new representation for task-oriented dialogue systems that models dialogues as dataflow graphs. The substantially improved expressiveness makes it natural to represent highly contextualized multi-turn, cross-domain dialogues that are difficult, if possible, to represent with existing dialogue representations such as intents and slots. This is the technical backbone of the newly released conversational interface for Microsoft Outlook.
2. Coverage. I will describe a recent effort in developing conversational AI agents that can simultaneously support hundreds of different domains and discuss the non-i.i.d. generalization challenge naturally arising in this setting.
3. Learning from use. If time permits, I will also discuss a new paradigm for building conversational AI agents where we empower the agent to proactively interact with users to resolve its uncertainties. In doing so it also gets to accumulate targeted training data to continuously and autonomously improve itself over time. This reduces the cost and privacy risks of conversational AI development while improving its explainability and trustworthiness.

Speaker Bio

Yu Su is an Assistant Professor at the Ohio State University and a Senior Researcher at Microsoft Semantic Machines. He got his Ph.D. from University of California, Santa Barbara and his bachelor’s degree from Tsinghua University, both in Computer Science. The overarching goal of his research is to democratize the access to data, knowledge, and intelligence through foundational and applied AI innovations, with a recent focus on conversational AI, knowledge base construction and reasoning. His work at Microsoft has led to a new conversational interface for Microsoft Outlook. He has published over 30 papers at premier AI conferences and regularly serves on the organizing and program committee of these conferences. His research has been recognized with awards such as Outstanding Dissertation Award from UCSB and Best of IEEE ICDM Selection 2019.