Scientist collects brainless blogs

Hey there, time traveller!
This article was published 22/03/2010 (4757 days ago), so information in it may no longer be current.

For Andrew Gordon, there’s no such thing as a boring blog — even if it chronicles making breakfast or walking to work.

A research scientist at the University of Southern California’s Institute for Creative Technologies, he’s heading a new project with the ambitious aim of archiving every English-language blog entry posted online — a million of them a day — in hopes of using this vast database to teach artificial-intelligence computers about real life.

“People write about the mundane aspects of their daily life and for me, personally, I find it incredibly interesting,” he says.

“Whereas I’m not particularly interested in people’s cats, I’m extremely interested in the experiences people have, how they perceive those experiences in the real world and then how they narrate those perceptions.”

Gordon’s work is part of a growing field of research that’s just starting to mine the massive quantity of thoughts, feelings and experiences real people pour out daily on the Internet. As the number and variety of bloggers has exploded — 130 million around the world, according to search engine Technorati’s latest estimates — they closely mirror the population at large, Gordon says, and advances in “natural language processing” allow researchers to parse their writing in increasingly sophisticated ways.

“The opportunity there is one that we’ve never had before as a research community,” he says. “I think it will radically change a lot of our theories, a lot of our understanding of social life and daily life.”

He believes this “narration of the mundane” is the best way to teach computers the common sense knowledge about everyday life that comes naturally to humans but is extremely difficult for artificial intelligence to grasp.

In the research world, the classic example of the difficulty in helping machines understand real life is the “egg-cracking problem,” Gordon says. If you told another person you smacked an egg against a bowl, they could instantly infer what would happen and the result if the bowl were made of cardboard or the egg was only lightly tapped against it, he says. But without knowledge about everyday life, a computer would be flummoxed.

“It’s this common-sense prediction that people are really good at and computers are really bad at,” he says.

Researchers have been working on artificial intelligence for 50 years, Gordon says, but no one has solved the problem of how to share this enormous store of information with the machines. As a young researcher, he used to think they’d recruit “really smart people” to hand-craft the rules, he says, but there simply aren’t enough scientists to feed computers everything they need to know. So why should humans want computers that can talk to them and understand their world?

Video-game designers are already creating “virtual human” characters, Gordon says, but when you stroll past them in a digital world, they don’t have the intelligence to interact.

“They look beautiful, but their brains are rather empty,” he says.

— Canwest News Service