Comp Semantcs Info Extraction

3 Credits · Level: Graduate, Undergraduate

This course focuses on computational issues in lexical semantics, with an emphasis on linguistic aspects of finding information in collections of documents. The underlying theme is the application of vector space representations, which have recently attracted renewed interest and research. We'll examine various problems connected with this theme, including idioms and collocations, lexical ambiguity and sense extension, document categorization, topic modeling and detection, and information extraction tasks. There will be several small-scale assignments, and students will jointly complete a project to be decided on once the class starts. Some familiarity and comfort with programming, mathematics, and statistics, in addition to linguistic concepts, is therefore desirable.