Example Student Research Projects
Carolyn Hafernik (’06) was interested in helping travelers find useful information for planning trips and for spur of the moment travel.
Carolyn’s thesis work explored ways to disambiguate geospacial references. One of her approaches automatically identified items such as place names (e.g.
Bangor, Maineversus Bangor, Ireland), name variations (e.g. , Bay Area, the Bay, SF), and landmarks relevant to a specific query. San Francisco
Desislava Petkova (’05) explored an approach that could help physicians diagnose disease.
In Desi’s thesis work, she developed clustering algorithms for automatic image annotation.
Problem: Given a large collection of images (for example, your vacation photos or all the photos you take in your college career), how can a system automatically label each image with a descriptive word or words that accurately describes what is pictured? Desislava’s system analyzed medical images such as x-rays and MRI’s and automatically assigned medical descriptors to each image.
Desislava published a paper on her work and presented it at the CLEF ’04 workshop in
Preeti Agrawal (’03) and Lin Gong (’03) worked on a system to translate English to Spanish.
Preeti and Lin explored the issues and problems associated with Cross-language Information Retrieval (CLIR). Given a query in one language (e.g. English), CLIR systems return relevant documents in other languages (e.g. Chinese, French). They used machine readable-dictionaries (MRD’s) to bridge the language gap between queries and documents written in different lanugages. They found that translation quality is greatly impacted by factors such as the quality (size, coverage, generality, etc) of the MRD, and the ambiguity of human language (Merriam-Webster’s online MRD lists 18 different meanings for the word bank; e.g. ‘bank’ could mean ‘riverbank’, a financial institution, a row of items, to drive into a cushion (as in pool), etc).
Last updated 8/2008