A sure sign of brilliance is tricking others into having fun while unwittingly helping you do work.
Carnegie Mellon University computer scientist Luis von Ahn was recognized for achieving that feat Tuesday when the October issue of Popular Science hit newsstands, naming him one of its 5th annual "Brilliant 10."
The list recognizes young, relatively unknown scientists who are breaking new ground.
"Yeah, I don't know," said von Ahn, 27, of Shadyside, with embarrassed modesty, when asked how it felt to be labeled "brilliant" at such an early stage in his career.
Von Ahn, a native of Guatemala, is leading research at CMU to develop tests that determine whether visitors to Web sites are humans or software robots deployed by spammers.
He is perhaps best known for creating the popular "ESP game," which harnesses brainpower to label the millions of images on the Internet.
"There's a lot of things that computers cannot do yet -- very simple things like understanding images," von Ahn said. "I get people to do these things for computers by playing games that also solve problems that computers cannot solve for themselves."
Players of the ESP game are paired anonymously. After being shown random Web images, the players type one-word descriptions. If the descriptions match, a label is born. Since the game's launch three years ago, more than 100,000 users have labeled at least 40 million images, von Ahn said.
The labels are particularly useful when Internet users search the Web for images. If images aren't labeled, search engines have a hard time finding them.
Google Inc. recently licensed the technology from CMU and launched a similar, experimental application last week called Google Image Labeler.
Von Ahn, who earned a doctorate in computer science from Carnegie Mellon in 2005, continues to develop online games that enlist people's reasoning skills. His latest, Phetch, asks players to write descriptive image captions that could make the Internet more friendly to the visually impaired.
"Basically, I want to have millions of people come and do useful things for humanity while they play games," von Ahn said.
The next challenge is to begin using data collected from these games to develop mathematical formulas that teach computers to recognize images on their own, von Ahn said. Improved computer vision would have powerful applications, such as monitoring security cameras with less manpower or helping doctors interpret medical scans.
Von Ahn is the second Carnegie Mellon computer scientist to be named part of Popular Science's "Brilliant 10" in two years. The magazine honored former CMU assistant professor Doug James, now at Cornell University, last fall, for developing tools that simulate collisions between virtual objects with greater speed and accuracy.
Larger text
Smaller text
Print this article
E-mail this article
