[Colloquium] Guest Speaker Announcement

[Ponda Barnes]

Guest Speaker
 
Presented by: Toyota Technological Institute at Chicago
 
Speaker: Alexander Morgan
Speaker's home page:
http://www.udri.udayton.edu/NR/exeres/14F9E30B-FF91-48A2-AB4E-C668831ADD7D.h
tm
 
Date: Wednesday, June 20, 2007
Time: 11:00 am
Location: TTI-C 1427 E. 60th Street, 2nd Floor
 
Title:  Information Extraction for Ungrammatical Text: 
Using a Part-Category Ontology to Index Repair Technicians' Notes
 
Abstract:
 
It is useful to be able to index repair notes by symptom, because a number
of decision-support systems rely on accurate symptom information.  Early
warning systems look for unusual symptoms or symptoms that occur with
unusual frequency.  Textual case-based reasoning systems depend on
smart-search on symptoms, made much more effective with pre-computed
indices.  Pareto reports on most-frequent symptoms are useful for
understanding problem areas. 
The most common sort of symptom simply indicates that a part is not
functioning properly.  Other symptoms involve smells, noises, handling
problems, and a few other categories.  The part-related symptoms are the
most difficult to categorize, because there are so many types of parts.  The
other symptoms can be handled by a few special-purpose ontologies.  This
project focuses on identifying which types of parts are mentioned in a block
of free text in a repair technician's note. 
Part mentions need to be identified in the text.  Because the text is
"ungrammatical," one cannot always rely on technology such as noun-phrase
extractors for this.  The part mentions must be matched with known
categories of parts and then disambiguated.  "Gas" does not always mean
gasoline.  It might mean the gas used in air conditioning.  This talk will
describe the information extraction process in general, and then on the ways
that ontologies can be used. 
 
Bio:
 
Dr. Alexander Morgan has been a research scientist at General Motors for
more than 25 years. His main areas of research include the numerical
solution of systems of polynomial equations and the development of practical
knowledge systems.  Most recently, he has been involved in projects
involving data mining, text analysis, and information extraction for health
care, quality, and warranty databases.  He has a Ph D. in mathematics from
Yale University 
 
 
If you have any questions or would like to meet the speaker, please contact
Ponda Barnes at pondabarnes at tti-c.org
For future TTI-C talks and events, please visit
http://ttic.uchicago.edu/cal/month.php
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20070619/d604a54d/attachment.html