By John Cosgrove


In Guatemala in 1986, not many people had computers—especially eight-year-olds. So when Luis von Ahn's mother gave him a Commodore 64, he had no idea how to use it.

But he was excited about the games on the computer, and through long days of trial-and-error, he mastered the Commodore's workings. He even figured out how to create his own programs. Along the way, he launched a career.

Von Ahn came to the United States to major in math at Duke, where he graduated summa cum laude. He chose Carnegie Mellon for graduate studies in computer science, receiving a prestigious Microsoft Research Fellowship. Today, at 28, he's an assistant professor whose work has a direct impact on users of the World Wide Web.

Here's an example. In 2000, Yahoo's chief scientist called Carnegie Mellon for help. Every day, major Web sites were experiencing millions of bogus registrations generated by computers. Yahoo needed a way to make sure that a person was logging in, not a computer program. Von Ahn was a member of the Carnegie Mellon team (led by professor of computer science Manuel Blum) that responded. The team developed Captcha, a feature that requires the user to identify distorted letters and numbers before logging in. It essentially eliminates fake computerized logins. Von Ahn estimates that it's used as many as a hundred million times a day on thousands of Web sites.

Von Ahn also developed a project to address the fact that computer programs can't recognize the content of images. That makes it difficult, for instance, for a search engine to provide a complete response when a user asks it to find particular pictures. Von Ahn's inspiration was to put human brains to work on the problem by exploiting people's love of games. His ESP Game (www.espgame.org) asks two online players to identify a series of images; they earn points for agreeing on their descriptions.

Google soon heard about the ESP Game and licensed it (renaming it Google Image Labeler) to enhance its image-search capability. Von Ahn estimates that, with heavy user participation, all the images on the Web—billions of them—could be labeled in just two months.

The ESP Game exemplifies what von Ahn calls "human computation"—harnessing human power to do jobs that computers can't. He says, "Tens of thousands of people worked on projects like the pyramids or the Panama Canal. Now, through the Internet, we can combine the efforts of hundreds of millions of people—you just have to trick them into producing something useful by doing things they'd be doing anyway."

Von Ahn's unique approach to problem solving has attracted attention. He recently was named one of Popular Science magazine's Brilliant 10 Scientists and received a MacArthur Foundation "genius" grant of $500,000. His only plan for spending the grant money is that it should be used "to do good," which he plans to do at Carnegie Mellon. He has received job offers from around the country—but he and his fiancée have no plans to leave Carnegie Mellon. (That's Laura Dabbish, who collaborated with von Ahn on the ESP Game and is an assistant professor at the Heinz School.) "This is a top computer science school, and I'm happy here," says von Ahn. "I also want to continue working with Manuel Blum, and Carnegie Mellon is the best opportunity for Laura."


RELATED LINKS:
Luis von Ahn
Microsoft Research Fellowship
CAPTCHAs
ESP Game
Google Image Labeler
Brilliant 10 Scientists