MIT EECS Undergraduate Research and Innovation Scholar
Clustering Related Word Senses in ConceptNet
ConceptNet is a network linking semantics, translations, and other aspects of words according to their lexical relationships. Word senses in ConceptNet refer to different meanings of a word. Computers can use word senses and their relationships to find information and understand peoples goals. ConceptNet collects senses from a number of different sources, including Wikipedia and WordNet. My work aims to build an algorithm that can automatically cluster related word senses in ConceptNet. The clustering will increase the coverage of senses in a complementary lexical resource. Enhanced information about clustered senses can be obtained combining the strengths of various resources.
I worked with Prof. Berthold Horn to develop a model using Java to simulate a method for suppressing traffic flow instabilities. I worked at Oracle on implementing a web application that allows users to query customer information and obtain tabulated results.