Title: Knowledge in Natural Languge Processing
Speaker: Lev-Arie Ratinov (UIUC)
Time: 2pm-3pm, Thursday, Nov 10
Place: Room C203, CUNY Graduate Center, 365 Fifth Ave (34str&35str).
		

Abstract:

In the past decade the importance of natural language text processing
(NLP) has grown immensely. The first steps in NLP applications involve
identification of topics, entities, concepts, and relations in text.

Traditionally, statistical models have been successfully deployed for
the aforementioned problems. However, the major trend so far has been:
scaling up by dumbing down- that is, applying sophisticated
statistical algorithms operating on very simple or low-level features
of the text. This trend is also exemplified, by expressions such as
``we present a knowledge-lean approach'', which have been
traditionally viewed as a positive statement, one that will help
papers get into top conferences.

In this talk I am making an argument that it is essential to use knowledge
in NLP, propose several ways of doing it, and provide case studies
on several fundamental NLP problems. I'm planning to give this talk in
two parts at CUNY. In part 1, I'll be covering using Wikipedia
knowledge for text classification and using knowledge to overcome data 
sparsity in named entity recognition. In part 2, I'll be covering 
disambiguating named entities to Wikipedia and using this knowledge to 
improve co-reference resolution systems.

Speaker's Bio:

Lev-Arie Ratinov is a Phd student at University of Illinois at 
Urbana-Champaign in the Cognitive Computation Group. His advisor is 
Dan Roth. His research interests are in semi-supervised learning, 
domain adaptation, sentiment analysis and using encyclopedic/external 
knowledge for the above tasks. He is also interested in applying 
Machine Learning techniques to many domains, such as natural language 
processing, finance, medical, information retrieval.