My current interest is in statistical natural language processing using semi-structured web resources. Randomly grabbing websites and processing them for information is fraught with problems -- issues such as web formatting, content checking and poor grammar cause current techniques to be rather limited and "shallow" in scope. However, semi-structured resources such as Wikipedia can eliminate some processing issues and enable learning of similarity metrics or information content on the given data set. This, in turn, may be utilized to create more intelligent agents, search engines, advertising targeting or any other application where real-world information may benefit performance. Additional interests include the development of agent technologies for embodied dialogue understanding. Basically, I want to have computers understand the "real world" situations where the spatial organization of the entities matter. So, you should be able to say to a computer "the one on the right", and it would understand what you are talking about. I'm co-advised by Dr. Eric Fosler-Lussier and Dr. Chris Brew. I've done extensive work with Dr. Donna K. Byron and have participated in research with Dr. Hui Fang and Dr. Craige Roberts. |
Research Reports |
|
Presentations |
|
Class Handouts |
|