University of North Texas Researchers Are Google-Eyed

By Aman Batheja, Fort Worth Star-Telegram, Texas

Nov. 2–Work by a researcher at the University of North Texas could someday out-Google Google when it comes to quickly searching for information online.

So it’s no surprise that the Internet search giant has given Rada Mihalcea a grant, valued at $107,112, to continue her research on information retrieval.

It may be pocket change for a company valued at $102 billion, but for Mihalcea, 31, an assistant professor of computer science and engineering, the gift will probably support her research for the next two years, she said.

Since 1997, Mihalcea has studied the field of natural-language processing, which looks at how language is used in information technology.

Her research focuses on extracting information from texts online. Mihalcea said she’s working on creating a program that could sift through large texts online and create key-word indexes and short summaries of the texts.

“There are zillions of news articles and millions of books. There’s just too much information,” Mihalcea said. “If [people] have access to succinct information, I think that could be better.”

Mihalcea’s research group at UNT has already developed technology that can summarize short texts such as news articles. The challenge with larger texts comes from creating a program that can determine what’s relevant to the text’s larger themes, she said.

That could have implications for Google’s fledgling Google Print endeavor.

The company announced last fall its intention to scan the world’s books and make them available online. Thousands of books in the public domain and books that publishers have submitted to Google are already available on print.google.com. Google’s plan is for copyrighted books to be searched as well but only allow short passages that mention the search key word to be available online.

Last December, Mihalcea’s work in natural-language processing earned her an invitation to speak at Google’s headquarters in Mountain View, Calif. Mihalcea spoke about her research in Google’s pioneering Pagerank algorithm, the foundation of its search engine. The title of her speech was Using Your Algorithm to Solve My Problems.

She received the Google grant in September.

Mihalcea said she doesn’t know how her work will ultimately be used.

The ability to quickly summarize large texts has wider-reaching implications than the uses of any one search engine, she said.

“It could change the way people access information.”

—–

To see more of the Fort Worth Star-Telegram, or to subscribe to the newspaper, go to http://www.dfw.com.

Copyright (c) 2005, Fort Worth Star-Telegram, Texas

Distributed by Knight Ridder/Tribune Business News.

For information on republishing this content, contact us at (800) 661-2511 (U.S.), (213) 237-4914 (worldwide), fax (213) 237-6515, or e-mail reprints@krtinfo.com.

GOOG,