I couldn't download this one; it required an account.  
http://www.upi.com/NewsTrack/view.php?StoryID=20060925-055616-9311r

There were 12 comments on this as of Sept 26, 2006

NewsTrack - Science
Computers taught to sort opinion from fact
WASHINGTON, Sept. 25 (UPI) -- The U.S. Department of Homeland Security
is funding a research project designed to teach computers to scan text
and then sort opinion from fact. 

The project involves Cornell University Professor of Computer Science
Claire Cardie and associate professors of computer science Janyce
Wiebe of the University of Pittsburgh and Ellen Riloff of the
University of Utah. 

The consortium is one of four University Affiliate Centers to conduct
research on advanced methods for information analysis and to develop
computational technologies that contribute to national security. 

"Lots of work has been done on extracting factual information -- the
who, what, where, when," explained Cardie. "We're interested in seeing
how we would extract information about opinions." 

The scientists will use machine-learning algorithms to give computers
examples of text expressing both fact and opinion and then teach them
to tell the difference.