Ed Lazowska explains it as an embarrassment of information riches. A decade deep into the twenty-first century, we’re slathered in data—from traffic cams, sensors on the ocean floor, satellite telemetry, your iPhone lens aimed at a kitten. “We’ve gone, rather quickly, from a data poor to a data rich society,” says Lazowska, who helms the University of Washington’s eScience Institute. The Northwest Institute for Advanced Computing—a new collaboration between Lazowska and his UW colleagues and the U.S. Department of Energy’s Richland-based Pacific Northwest National Laboratory—will tackle the global glut of info. Via “machine learning,” or programs that “learn” on their own, the team will develop computational methods that sift through data and determine which is relevant for a particular line of research. Back in the day, Lazowska says, “a grad student would spend three years building a sensor of some sort, then two weeks freezing on a mountaintop, and come back with hardly enough data to justify all that time and effort.” Now that the data itself is the mountain, Lazowska and his new Department of Energy partners hope to bring us all in from the cold.

 

Published: April 2013