MIT researcher aims to understand language with Human Speechome Project

It's far from the first time a researcher has enlisted the help of his own family or kids, but MIT's Deb Roy's latest endeavor looks to be a bit more ambitious than most, as he's aiming to do nothing short of understand how children learn language. To do that, Roy and his wife installed 11 video cameras and 14 microphones throughout their house to record just about every moment of their son's first three years. That, obviously, also required a good deal of computing power, which came in the form of a temperature-controlled data-storage room consisting of five Apple Xserves and a 4.4TB Xserve RAID (you can guess why Apple's profiling 'em), along with an array of backup tape drives and robotic tape changes (and an amply supply of other Macs, of course). While the project is obviously still a work in progress, they have apparently already developed some new methods for audio and video pattern recognition, among other things, and it seems they'll have plenty of work to sift through for years to come, with the project expected to churn out some 1.4 petabytes of data by the end of year three.

[Thanks, Jeff]