Invited talk at ICT
I gave a talk titled "Asynchronous Context Management
for Embodied Agents" at the Institute for Creative Technology, University of
Southern California. The slides are available to download. 
September 21, 2007
New 3-year NSF grant!
"Establishing and Breaking Conceptual Pacts with Dialog Partners": I will be working with Joy Hanna at Oberlin College to see if people interacting with computer-based dialog partners form the same representations of common ground as they do when interacting with human partners. The question is important for designing the referring expression generation component of software that interacts in natural language.
September 7, 2005
Invited Talk at Northwestern University's Articulab
"Context Management for Embodied Dialog Agents"
August 5, 2005
HLT demonstration program announced
The demos to be presented at HLT/EMNLP2005 have been selected. The list is available on the conference website under the 'demos' tab.
August 2, 2005
Look-ahead speaker dates for the upcoming schoolyear
The following speakers have been lined up to visit OSU next year as part of the Pragmatics Initiative. More will be posted as their dates are confirmed.
- October 13 and 14, 2005, Jason Stanley from Rutgers.
- Nov. 3 and 4, 2005, Jerry Hobbs from the Information Sciences Institute at USC.
- January 26, 2006, David Beaver from Stanford.
- February 23, 2006, Chris Potts from UMass Amherst.
Aug 1, 2005
Publications page updated with recent results. Here are some highlights:
Donna K. Byron and Laura Stoia. An analysis of proximity markers in collaborative dialog. In Proceedings of the 41st annual meeting of the Chicago Linguistic Society. Chicago Linguistic Society, 2005.
Abstract: Why do speakers moving through a virtual world use the close-marked phrases "this/here" vs. the alternative "that/there"? Is this choice purely based on their spatial distance from an object? We find three factors that bear on this choice: space, time, and task obligations.Sarah Brown-Schmidt, Donna K. Byron, and Michael K. Tanenhaus. Beyond salience: Interpretation of personal and demonstrative pronouns. Journal of Memory and Language, 53(2):292-313, August 2005.
Abstract:A series of studies on how listeners process the pronouns it and that differently and how they respond when instructed to "Put that..." vs. "Put it...". When we kept linguistic factors such as discourse salience and sentence structure constant, but changed properties of the spatial configuration of the objects to be manipulated, subjects changed their interpretations of the pronouns.(if you have a ScienceDirect subscription)
Donna K. Byron, Thomas Mampilly, Vinay Sharma, and Tianfang Xu. Utilizing visual attention for cross-modal coreference interpretation. volume 3554/2005, pages 83-96, 2005. Springer Lecture Notes in Computer Science: Proceedings of Context-05.
Abstract:In this study, we developed an algorithm to keep track of this items a speaker is looking at over time, in order to predict which object the speaker would talk about next. The method is offered as an alternative to using the discourse history to predict the topic of an upcoming sentence. TheAdriane Boyd, Whitney Gegg-Harrison, and Donna Byron. Identifying non-referential it: A machine learning approach incorporating linguistically motivated patterns. In Proceedings of the ACL Workshop on Feature Engineering for Machine Learning in Natural Language Processing, pages 40-47, Ann Arbor, Michigan, June 2005. Association for Computational Linguistics.
Abstract: English can include expletive it, which has the same form as a pronoun but is not referential. These items need to be ignored during pronoun resolution instead of being resolved as though they were referential pronouns. In this study, we developed a machine learning algorithm to automatically detect expletive it in the BNC sampler.
June 26, 2005
ACL Workshop on Teaching Computational Linguistics and NLP
I was a Panelist for a discussion on teaching natural language processing and computational linguistics at the ACL05 workshop on teaching. The main points I wanted people to take away from my comments were:
- Encouraging undergrads toward a career in CL research is sometimes accomplished with very simple things. If you teach a CL or AI class, make sure the people who did well in the class know they did well. If possible, post your grade distributions so the students at the top can see how well they did. This is especially helpful for under-confident students.
- In my Comp Ling class, I set aside one class period a week to have students
interact with a variety of working CL technologies so that
they can see a broad spectrum of products related to language analysis
and production. These sessions let the students get hands-on
exposure to concepts or techniques that we didn't have time to explore in
homework assignments. I copied this idea from Eric Fosler-Lussier's
speech recognition class, and I thought it was very effective, modulo
technical snafus. The technique works best with software that can be
loaded to student laptops before class starts. My play-day instructions are available
in the course web page in case anyone wants to use them. In SP05, the
in-class exploration covered:
- limitations of Eliza-style chatbots
- Probabilistic tagging and parsing with the Bickel parser
- Semantic tagging:Named entity tagging, semantic roles, and Framenet
- 5 different styles of talking heads (veepers, sitepal, Mary101, peopleputty, and the facegen demo movie)
- dialog-act tagging with the DAMSL tagset
June 7, 2005
May 17, 2005
Brownbag talk in OSU Psychology Dept.
I gave an overview of my research to the Ohio State University
Psychology Department in their brownbag speaker's series. The title
was "Enhancing Computational Models of Pronoun Resolution." The
slides are available to download.
May 5, 2005
OSU Lima Campus Women in Science and Engineering
I spoke to the Lima Campus's women in Science and Engineering group,
organized by Sabine Jeschonnek,
and introduced them to the field of Computational Linguistics and CL
research. The talk was titled "What is Computational Linguistics,
and what does a Computational Linguist do?". I had a really fun time
meeting the group, learning about Physics
from Sabine, and telling them about CL. They were a really great
audience!