|Location||EB3 rm 2213|
|Date||March 25, 2011 1:00 PM|
In the last two decades, Natural Language Processing (NLP) has almost exclusively worked on English. We know how to analyze English, how to produce it automatically, and how to translate into English. However, when we apply the methods developed for English to other languages, at best, the performance is significantly below the results for English, but in many cases, the methods cannot be applied at all. In this talk, I will give an overview of the research problems that researchers in NLP investigate at current. Then, I will focus on three problems, syntactic parsing, coreference resolution, and machine translation, and I will show the problems that we encounter if we use state-of-the-art methods for other languages.
Sandra Kübler is an Assistant Professor and the Director of Computational Linguistics in the Department of Linguistics at Indiana University. She holds a PhD in Computational Linguistics from the University of Tübingen and an MA from the University of Trier, both in Germany. Before she moved to Indiana, Kübler spent time at the University of Tübingen and the University of Duisburg. Her main research areas are dependency parsing, parsing for morphologically rich languages (mostly German), and machine learning for Computational Linguistics problems (e.g. for anaphora resolution or word sense disambiguation). She is interested in how to integrate linguistic information into machine-learning approaches to Computational Linguistics.