Pittsburgh Innovates

What does having the latest Internet news and gossip at your fingertips have in common with the location of contamination in a water supply system?

Almost nothing, actually, except that both can be determined with a versatile algorithm developed by CMU researchers. Using a problem-solving method called Cascades, Carlos Guestrin, CMU assistant professor of computer science and machine learning, has compiled a list of the best 100 blogs in the blogsphere, sites deemed to be on the cutting-edge of news.

There are the well-known blogs, such as Instapundit and Boing Boing, but also some more obscure ones like Watcher of Weasels. To see the list, click here.

The biggest surprise was that sites like slashdot.com and CNN.com didn’t make the list, says Guestrin. “But slashdot is a big blogsite that generates stories that don’t go anywhere. The cascade picked up the big buzz stories. Reading these sites, you’ll be on top of everything without having to do as much work.”

“The cascade is pretty intuitive,” he adds. “You post a story, somebody points it out, passes it along and it generates a buzz. In a water contamination setting, the spreading thing works the same way. Contamination spreads to neighboring pipes the same way the a story would spread from one person to another.”

The report on the blog and water system case studies, “Cost-Effective Outbreak Detection in Networks,” was presented at the Association for Computing Machinery’s International Conference on Knowledge Discovery and Data Mining earlier this year.