Marilyn Walker, is a Professor of Computer Science at UC Santa Cruz,and a fellow of the Association for Computational Linguistics (ACL), in recognition of her for fundamental contributions to statistical methods for dialog optimization, to centering theory, and to expressive generation for dialog. Her H-index a measure of research impact is 55. Her current research includes work on computational models of dialogue interaction and conversational agents, analysis of affect, sarcasm and other social phenomena in social media dialogue, acquiring causal knowledge from text, conversational summarization, interactive story and narrative generation, and statistical methods for training the dialogue manager and the language generation engine for dialogue systems. Before coming to Santa Cruz in 2009, Walker was a professor of computer science at the University of Sheffield. From 1996 to 2003, she was a principal member of the research staff at AT&T Bell Labs and AT&T Research, where she worked on the AT&T Communicator project, developing a new architecture for spoken dialogue systems and statistical methods for dialogue management and generation. Walker has published more than 200 papers and has 10 U.S. patents granted. She earned a B.A. in Computer and Information science at UC Santa Cruz, M.S. in Computer science at Stanford University, and a M.A. in Linguistics and Ph.D. in Computer Science at the University of Pennsylvania.
Her Program :
- Tuesday, july 3 2018 - 2:00pm - LIMSI, conférence room - Rue John von Neumann, Orsay - Plan Accès
- Title : Stylistic Variation in Natural Language Generation for Dialog
- Abstract : When humans converse, the style of their utterances reflects both their own individual differences and aspects of the situation, such as whether the situation is formal or informal. Stylistic variation is also intro- duced when people adapt their utterances to their conversational partner, as can be commonly perceived when someone talks to a child. It has also been shown that humans adapt to a dialogue system just as they would to a human partner. However it is much more challenging to engineer a dialogue system to adapt to the human. I will discuss techniques for generating stylistic variation for dialogue using two different methods. The first method enhances a statistical language generator with additional parameters for generating personality-based stylistic variation, and then learns to set these parameter values to manifest particular per- sonality types. The second method harvests paraphrases from user reviews in a domain, such as restaurants, to learn different ways to say the same thing, and support, for example, variations in expressiveness related to hyperbole.
- Friday, july 6 2018 - 9:30am - Jade amphi, telecom-paristech, 46 Rue Barrault, 75013 Paris, Metro Corvisart/Placed'Italie
- Title: Identifying High Quality Arguments and Argument Facets in Social Media Dialogues - All information
- Tuesday, july 17 2018 - 2:00pm - LIMSI, conférence room - Rue John Von Neumann, Orsay - Plan Accès
- Title : SlugBot: UCSC’s Experience in the Alexa Prize Competition
- Abstract : UCSC was one of the 12 funded teams to compete in the 2016-2017 Alexa Prize Competition funded by Amazon. The goal of the competition was to create an open-domain conversational agent that could talk about any topic and carry on a conversation for at least 20 minutes. This lecture describes the basic architecture of Slugbot, technical innovations used in SlugBot, and the limitations of current state-of-the-art technology that make it difficult to creat a truly conversational agent. We focus on four key underlying functionalities that need further research: (1) general ontological and common-sense resources; (2) methods for the integration of structured and unstructured content resources; (3) discourse management and discourse state tracking; (4) evaluation metrics and methods.