Computer Science Colloquium, 2004-2005

Ed Hovy
Information Sciences Institute, University of Southern California
November 24th, 2004

Ontological Modeling of Information-Rich Organizations by NLP Analysis of Email

Over the past few years, Social Network Analysis has developed several innovative methods for inferring facts about the nature of organizations by examining patterns of interaction among their members. This talk describes what one can infer using Natural Language Processing (NLP) techniques for the automated analysis of emails. Focusing on two principal dimensions -- topics and speech acts -- the research explores how one can learn quite surprising things about expertise, projects, social and organizational roles, etc.

Shuly Wintner
Last modified: Fri Oct 22 07:30:52 IST 2004