Saiyam Kohli under Dr. Ted Pedersen, Computer Science
During the Summer of 2005, Computer Science graduate student Saiyam Kohli worked on a VDIL sponsored project entitled "Visualizing Relations Between Concepts in WordNet". This project was supervised by Dr. Ted Pedersen, Associate Professor of Computer Science.
WordNet is a freely available lexical database that links together sets of synonyms according to a variety of semantic relations, including 'is-a-kind-of', 'is-a-part-of', and 'is-a-way-of-doing'.
As such WordNet provides information about concepts and how they are related to each other. Simple examples of the kind of information available include the following: a cat is a feline, a feline is a carnivore, a steering wheel is a part of a car, a car is a vehicle, to walk is a way of moving, etc. There are more than 100,000 concepts in WordNet, connected by nearly 200,000 relations, so it clearly forms a very large network of concepts and relations that is impossible for a human to mentally picture or manage.
The goal of this project was to create a visualization tool that would allow a WordNet user to easily see relations between concepts, especially those among concepts that are not directly linked within WordNet. To do this we added a visualization component to the WordNet-Similarity package which has previously been developed by Dr. Pedersen and former Computer Science graduate students Siddharth Patwardhan (M.S., 2003) and Jason Michelizzi (M.S., 2005).
This summer Mr. Kohli implemented visualizations of a number of well known measures of similarity and relatedness that are included in WordNet-Similarity. Of particular interest was the measure of Hirst and St-Onge (1998), which finds relations between concepts that are joined by paths of relations that have certain characteristic shapes and lengths.
This tool was implemented in Perl using the Gtk2 library, and runs on Linux systems. The software has been made freely available via the Comprehensive Perl Archive Network (CPAN), and can be downloaded from http://search.cpan.org/dist/WordNet-Similarity-Visual/.
Mr. Kohli spent approximately 60 hours working in the VDIL lab in the summer of 2005, and approximately 90 hours working outside of the lab on this project, for a total of 150 hours. Dr. Pedersen spent approximately 15 hours meeting with Mr. Kohli and evaluating his work. The total award of $2,000 was given to Mr. Kohli as a summer stipend. Dr. Pedersen’s time was not compensated.