Emory Report
January 28, 2008
Volume 60, Number 17

Google yourself
It could be important to your career – not to mention your personal life.

“Web search is the way that many other people learn about who you are,” says Eugene Agichtein.

Keeping track of your online search results could give you an opportunity to correct any wrong information about you on the Web. Fewer than half of regular Internet users have ever googled their own names, he says, adding: “In this case, vanity could be a virtue.”

   
Emory Report homepage  

January 28, 2008
Raising IQ of Web searches

By carol clark

Eugene Agichtein was born in Moscow in 1976 — a time and place when access to information was not taken for granted. He recalls that illegal books by Solzhenitsyn and Bulgakov came to his family’s home in the form of loose-leafed, mimeographed pages that were quietly passed among friends.

“Of course, I wanted to read them all,” says Agichtein, who immigrated to the United States with his family when he was 13.

In 2006, Agichtein joined Emory, where he is assistant professor in the Mathematics and Computer Science Department and director of the department’s Intelligent Information Access Lab. He recently received a “Beyond Search” award from Microsoft Research, in conjunction with Charles Clarke of the University of Waterloo.

The award — including $82,000 in funding and data from more than 100 million actual consumer Web searches — is the latest boost for what Agichtein describes as his life’s work: helping people get whatever information they need online, in ways that are both efficient and intelligent.

“How can we allow people to gain access to these areas of knowledge that used to be closed off to them? I think that improving Web search is a powerful way to change the world. It’s a key transformational technology,” he says.

Google, Microsoft and Yahoo are the big three names in search engines — Google alone fields 1.2 billion search requests each day. So how do the search engines sort through millions of Web pages and determine how to rank the responses to this vast number of queries? A sophisticated system of analyzing the content of the pages and the link structure of a Web page is one method, but it is far from foolproof.

“The Web is kind of a wild place. People are always trying to come up with ways to game the system,” says Agichtein, citing “click spammers” and “link spam farms” as two such strategies. “It can be difficult to pull out the true information from the noise.”

The experience of the person doing the searching, along with the person’s native language and culture, also must be considered by scientists working to advance search engine technology. Some searcher habits have become ingrained. For example, more than half of Internet users will click on the top link of a search engine response, whether it’s relevant or not, Agichtein notes.

He breaks down his research into three basic components: the content on the Web, the link structure, and what he calls “the third dimension” of the Web — the users.

One focus of the Intelligent Information Access Lab is analyzing actual user data, to discover patterns of behavior and how people find information in various contexts.

“It’s fascinating, delving into how the human mind works,” he says. “Exploring how millions of people create content and search for information on the Web provides a unique window into cognitive processes.”

Search personalization, coupled with advances in wireless handheld devices and biometrics such as eye-tracking, will further speed changes in the already rapidly evolving field of Web search, Agichtein predicts. “Ten years from now, computerized searches will look much different than they do today — you won’t be just typing words into a box on a screen,” he says.
|