We're one step closer to browsing the internet with our mind while our robot butler brings us drinks, and it's all thanks to Intel. PC Magazine tells us about the wonderful tech on display at the company's latest R&D day.
Behold HERB, the robotic centerpiece of the festivities at the most recent Intel R&D day, where new technology is shown off to the sort of folks who are keen to head back to their office and write about it.
HERB may look like a deadly mechanical torture device, but he's really just a robot butler, ready to serve your every need, unless that need involves killing other humans, I'm sure.
Developed by Quality of Life Technologies and Carnegie Mellon, HERB can pick things up, put things down, and carry things around. Impressive, but perhaps the most impressive thing about HERB (aside from his Segway base) is the glowing eye that acts as a laser scanner, allowing him to identify common household items, a task much easier for robots than humans, as evidenced by the next item on the list, the brain interface.
While robots can identify objects by sight, human beings often identify them with mental images, and these images can differ from one person to another. It's these complicated and diverse thought patterns that make it difficult for scientists to create an interface that reliably lets us bring up words using our minds.
But that's not stopping Intel Labs Pittsburgh researcher Dean Pomerleau from trying. In his quest to create a "Rosetta Stone" of brain activity, his team has utilized a $2 million magnetocephalograph to amass a vocabulary of more than 1,000 words, all mapped to the part of the brain stimulated by thinking about them.
Pomerleau pitted himself against the computer to demonstrate the technology, looking at brain scans to see if the test subject represented in it was thinking "cup" or "cabinet." The computer delivered the appropriate response just under 90% of the time.
Both of these are examples of some amazingly advanced technology, but for me, the real winner of the show is something called Meeting Diarist. Created by Nelson Morgan, a University of California professor and director and president of the International Computer Science Institute, Meeting Diarist solves a problem that has touched any journalist who's ever attempted to record and transcribe a roundtable meeting.
Meeting Diarist takes audio and attempts to use speaker identification to build a timeline of who was speaking when. Say you were in a crowd of people, and someone mentioned you being a dork. If you gave Meeting Diarist the audio, it could pick out the word dork, give you the time it was said, and if you had sufficient knowledge of who was talking in the first place, it could identify the speaker.
I need that technology in everything I own.
But that's just me. Hit up the link below to see if anything else from the Intel R&D day catches your eye, other than the sharpened prongs of HERB's helpful hands.
Intel's R&D Features Robot Butler, Brain Interface [PC Magazine]