Computer Science Research Seminar

 

Shane Bergsma
PhD Candidate, University of Alberta

Deep Linguistic Knowledge from Web-Scale Statistics

Monday, November 23, 2009
MC320 11:30am
Middlesex College

ABSTRACT:

At the scale of the World Wide Web, simple statistics can provide deep linguistic knowledge. In this talk, I discuss the impact the web has had on Natural Language Processing research. I describe several novel algorithms for extracting meaningful information from web-scale collections of text, and present systems that use this information to achieve superior performance. I introduce semi-supervised Machine Learning as an effective way to analyze web-scale data. The talk also touches on the evolution of the web in language research: from search engines to databases of phrase counts to futuristic Cloud Computing technology.

 

Refreshments to follow

 

Western provides the best student experience among Canada's leading research-intensive universities.