|
Microsoft to 'read' minds
18/01/2008 14:58 - (SA)
San Francisco - When a Microsoft Corp patent application for a method of sorting brain waves surfaced late last year, it drew quips that the company now plans to read PC users' minds, in addition to selling them software.
But the patent is only a starting point for a more broad - and benign - investigation into the subconscious, according to Microsoft researcher Desney Tan.
"It's a small piece of a larger project, with grander goals," said Tan, who leads a team focused on brain-computer interfaces.
Tan's team plans to present two more related innovations in the spring. The efforts are part of a general push by Microsoft to more intimately connect the human body with computers.
Building a better user interface is essential for Microsoft, as it competes with computing rivals such as Apple Inc. Microsoft Chairperson Bill Gates acknowledged as much in a speech earlier this month, when he said that of the elements needed for a new generation of computer applications, "One that people underestimate the most, I would say, is the power of natural user interface."
The companies' grappling over interfaces dates back to competition between early versions of Microsoft's Windows and Apple software in the 1980s.
Apple has made notable strides in this area, first developing the use of a mouse with a computer, and more recently developing a touch-screen wireless phone, the iPhone, which avoids the keypad and stylus common among rival devices.
Hoping to one-up its peers
But with new interfaces that can key on users' minds rather than a keyboard or touch screen, Microsoft hopes to one-up its peers in the future.
"It's our fundamental belief that computers are still relatively dumb, no matter how smart they act," Tan said. "We're trying to make computers smarter, and provide them with a mechanism to understand the user more deeply."
Following up on the initial brain sorting method described in last year's patent application, Tan's team is planning to present two more related innovations at an April conference.
One involves "cognitive load", Tan said, or "measuring how hard you're thinking in order to perform a task". By tracking that, a computer could "adapt itself in real time to what you're doing", Tan said, and "present you information so you learn it in an optimal manner".
The second innovation, Tan said, involves tapping a computer user's brain signals to rapidly sort and classify images as they appear, without the user even realising what he or she is doing.
Microsoft's research groups are relatively sheltered from the commercial needs of the company, and Tan isn't sure when his team's innovations might appear in a product. "Obviously we'd like to get this out to the mass market," he said.
Paul Sajda, an associate professor in Columbia University's biomedical engineering department, believes "it's going to be five to 10 years before it becomes clear what the viable business model is" for the brain-computer interfaces Microsoft is exploring.
While Sajda said it's logical for computing companies to explore the issue, Tan said he isn't aware of Microsoft's larger rivals undertaking similar research.
Apple representatives did not respond to a request for comment.
Google, a rival to Microsoft in the internet services business, has also begun offering software to compete with Microsoft's Office products, such as Word and Excel. A Google spokesperson said the company has nothing "to announce at this time" about research on brain-computer interfaces.
Leaky signals
Sajda said a better brain-computer interface could help people more efficiently search the web, "or wherever there is a lot of imagery". But an innovation will have to be significant in order for Microsoft to bother bringing it to market, he said.
"It's not going to impact the consumer market if all it's doing is decoding whether I want to press the right or left mouse button," he observed.
As more data is pushed at computer users in a deluge of messages and images, a brain-activated interface increasingly makes sense, Sajda said. The average computer interaction may soon be on par with that experienced now by air-traffic controllers or Wall Street traders, he said.
Tan's team is made up of a core of five researchers, with up to 15 working on related projects at any given time. While the expertise on brain waves has existed at Microsoft for some time, "it's just in the last two or three years that this has pulled into a coherent agenda," Tan said.
"Anything the human body does, we can probably measure," Tan said. Electronic signals from the brain become accessible, he said, when they leak through the skull and into the skin. "Based on that leaked signal, we can infer activity," he said.
A current hindrance is the need for users to wear an intrusive cap fitted with sensors, though Tan said he can envision a time when such caps will be unnecessary.
'We haven't even begun to scratch the surface'
Gaming should be a particularly receptive market for computer-brain interfaces, Tan said.
Closely-held Emotiv Systems Inc, based in San Francisco, has already demonstrated a system where video gamers wearing a headset can command the action on the screen using brain signals.
But such gaming applications are mere "gimmicks" compared to the benefits brain-computer interfaces could bring to a broader audience, Sajda said.
That broader audience is already becoming increasingly familiar with touch screen technology, an innovation that Tan thinks is a step along the way to brain-computer interfaces.
Apple's popular iPhone device features perhaps the most celebrated example of a touch screen yet, though Microsoft last year began promoting its own "surface" touch computing technology.
"I'm excited by seeing anything that challenges the status quo in computing, and certainly the touch stuff has done that," Tan said.
Though last year's patent brain signal application provoked unwelcome speculation about Microsoft's nefarious intentions, Tan said more related patents are likely on the way. "We haven't even begun to scratch the surface," he said.
Most importantly, Tan said, people should understand that the technology his team works with falls short of actually being able to read peoples' thoughts.
"A myth is that we can get what you're thinking," Tan said. "We can't, at least with current technologies."
|
