Department Seminar - by Dr. Ted Pedersen (UMN)
Updated on Tue, 01/29/2013 - 9:44pm
Measuring Semantic Similarity and Relatedness in the Biomedical Domain : Methods and Applications
Speaker: Dr. Ted Pedersen (UMN)
When: Fri, 04/19/2013 - 11:15am - 12:15pm
Room: CH 430
In this talk I will provide an introduction to techniques for measuring the similarity and relatedness of concepts found in biomedical ontologies and concept hierarchies. For similarity measures I will focus on methods that rely on shortest paths, depth in hierarchy, and information content. The discussion of relatedness measures will focus on methods that are based on the content of concept definitions, sometimes augmented with co-occurrence data from corpora. I will then describe how these measures can be used to tackle a variety of problems in natural language, including word sense disambiguation and sentiment analysis.
Ted Pedersen (PhD, 1998, Southern Methodist University) is a Professor in the Department of Computer Science at the University of Minnesota, Duluth. His research interests are in Natural Language Processing and Computational Linguistics, and focus on determining the meaning of words and phrases in written text. Recently he has been involved in developing an open source software package for measuring semantic similarity and relatedness among concepts found in the Unified Medical Language System (UMLS::Similarity). He previously led the effort to develop such a package for the general English lexical database WordNet (WordNet::Similarity). He has also done work on unsupervised discovery of word senses and supervised approaches to word sense disambiguation. He is the recipient of a National Science Foundation CAREER award and his research has also been funded by the National Institutes of Health.