CISC 882: Natural Language Processing

Catalog Description:
Introduction to computational models of syntax, semantics and pragmatics for natural language understanding. Emphasis on design of English interfaces to data bases and ill-formed input.

Current Texts:
Natural Language Understanding, Second Edition
James Allen
Benjamin/Cummings, 1995

Goals:
The course is intended as an in-depth survey of the field of natural language processing (NLP). While NLP must be concerned with both natural language understanding and natural language generation, to date most of the research has concentrated on understanding. We will therefore use natural language understanding as a vehicle to introduce the three major sub-fields of NLP: syntax (which concerns itself with determining the structure of a sentence), semanti cs (which concerns itself with determining the explicit meaning of a sing le sentence), and pragmatics (which concerns itself with deriving the implicit meaning of a sentence gained from it being used in a specific discourse context). The course will end with a look at some of these issues in the context of natural language generation.

Content:

• Parsing natural language
  • basic context-free grammars and parsing
  • augmented context-free grammars (and chart parsing)
  • handling movement: bounded and unbounded
  • some alternative formalisms: ATN's and Definite Clause Grammars
  • deterministic parsing
  • statistical methods
• Semantics
  • semantic sentence representation (logical form)
  • semantic interpretation
  • semantic grammars
• Pragmatic Processing
  • anaphora resolution/focusing
  • plan/goal recognition
  • discourse structure
• Introduction to natural language generation

Required Background: CISC 681 (Artificial Intelligence) or consent of instructor.