Answering complex questions

Although Google has become a verb for most of us, the search engine has its limitations: ask it for the impacts of the current global financial crisis in different parts of the world for example, and you can expect to sift through thousands of results for the answer.

Under the guidance of his supervisor, Dr. Yllias Chali, Sadid Hasan, a U of L Master of Science [1] (Computer Science [2]) student, is assisting in the development of a more intelligent way of searching the Internet – or any large collection of documents. Complex question answering, as it’s called, creates summaries of large volumes of information as answers to complex questions, rather than simply offering a listing of sources.

“We are researching a kind of search where we can ask questions and the answer will be available right away to the user,” explains Hasan, who came to the U of L from Bangladesh to work with Chali.

His master’s research project involves developing and testing software that applies complex question answering techniques to a collection of newspaper articles.

“I’m using machine learning techniques to generate summarization models.”

Chali has been working in this area for nearly a decade. Hasan, like all of Chali’s graduate students, was assigned a specific project related to his research program and carefully guided through the process.

“I try to give the students as much direction as possible to allow them to gain research skills quickly.”

Hasan is one year into his program but has already published a couple of papers and earned one of four $15,000 School of Graduate Studies Fellowships [3]. His success is the product of talent, but also tenacity, says Chali.

“Sadid has great research potential; he’s very hard-working.”