UW team part of IBM 'cognitive' computing chip project

Aug 19, 2011

(PhysOrg.com) -- University of Wisconsin-Madison researchers are part of the IBM-led team that has unveiled a new generation of experimental computer chips - the first step in a project to create a computer that borrows principles from a mammal brain.

The chips combine neuroscience and nanotechnology to operate more like by thinking and learning from . They also run on much less power than current technology.
 
UW-Madison neuroscientist Giulio Tononi leads the UW team, which is designing the software to teach the chips to learn and think.
 
"We are using the new IBM neurosynaptic to develop architectures that are good at integrating information - a key adaptive feature that the brain excels at, and which has proven difficult to achieve using conventional computers," says Tononi, professor of psychiatry at the UW School of Medicine and Public Health and an internationally known expert on consciousness.
 
The UW-Madison team includes professor Mikko Lipasti of computer sciences and scientists from both the computer sciences and the neurosciences. Along with Columbia University and IBM, they are the "software" team for the cognitive computer project.
 
Meanwhile, nanotechnology and supercomputing experts from Cornell and the University of California-Merced are collaborating with the IBM team to design the "hardware." Dharmendra Modha of IBM is the principal investigator of the project, which began in late 2008 and is called the Systems of Neuromorphic Adaptive Plastic Scalable Electronics (SyNAPSE) project.
 
Along with announcing the creation of the first two "cognitive" today, IBM also announced that the project has been awarded about $21 million in new funding from the Defense Advanced Research Projects Agency (DARPA) for the second phase of the project.
 
The project's ultimate goal is to create a small, low power-usage computer that analyzes complex information from multiple senses at once, but also dynamically rewires itself as it interacts with its environment.
 
The new prototype chips are a significant next step toward this goal. These chips each contain 256 digital "neurons" and won't be programmed the way traditional computers are today. Rather, cognitive computers are expected to learn through experiences, find correlations, create hypotheses, and remember - and learn from - the outcomes, mimicking the work of the brain.
 
The project leader, IBM's Modha, envisions small computers that can make decisions based on complex data. For example, a cognitive computing system in the ocean could use sensors that constantly record and report metrics such as temperature, pressure, wave height, acoustics and ocean tide, and issue tsunami warnings based on its decision making.
 
"We've managed to break through a number of intellectual and conceptual barriers by bringing together some of the best experts across neuroscience, nanoscience and supercomputing," says Modha. "This is an exciting time."

Explore further: Computer software accurately predicts student test performance

add to favorites email to friend print save as pdf

Related Stories

IBM pursues chips that behave like brains

Aug 18, 2011

Computers, like humans, can learn. But when Google tries to fill in your search box based only on a few keystrokes, or your iPhone predicts words as you type a text message, it's only a narrow mimicry of what ...

IBM and Samsung collaborate on chip research

Jan 13, 2011

US computer giant IBM and South Korean electronics titan Samsung have announced they will begin working together on ways to make better chips for smartphones and other gadgets.

Recommended for you

Ant colonies help evacuees in disaster zones

23 hours ago

An escape route mapping system based on the behavior of ant colonies could give evacuees a better chance of reaching safe harbor after a natural disaster or terrorist attack by building a map of showing the shortest routes ...

User comments : 5

Adjust slider to filter visible comments by rank

Display comments: newest first

Nanobanano
1 / 5 (1) Aug 19, 2011
I should contact IBM.

I know they don't deal in consumer goods so much, but I thought of a catchy consumer product that could be made with this technology. It could easily pay for the cost of the entire operation.
Nanobanano
1 / 5 (1) Aug 19, 2011
Man, hehe.

They could replace schools and teachers entirely ina few years, by using this technology to create computers which can teach the subject matter via access to the internet and compiling all relevant subject matter into an optimized learning program, AND answer a variety of student questions in ways that ordinary learning systems cannot be made to anticipate.

Imagine if you could replace teachers with a computer that costs only a few thousand dollars! The state governments and school boards would save so much money on education it's amazing. One teacher's salary for 1 year could pay for like 10 machines, and they'd each work for several years.

Heck, you might not even need schools at all any more!

Imagine the saves to the state governments and school boards! They would have so much money left over each year, they could cut taxes to near-zero. imagine all the money saved by not paying pensions to those over-paid, under-skilled retiring teachers!
FrankHerbert
1.7 / 5 (6) Aug 19, 2011
Are you serious? I'm sure every person working there has thought of this. I'm sure every sci-fi author has thought of this. I'm sure every fairly intelligent person has thought of this.

PLATO (http://en.wikiped...stem%29) has been around since the 60's. I used it in the mid-90's for algebra and it was great.

Unless you've actually made some breakthrough in AI, I doubt you have anything to contribute. I'll really start to worry if you begin copyrighting your posts.
Nanobanano
1 / 5 (1) Aug 19, 2011
Frank:

What you're talking about is a computer program in which the programmers attempted to anticipate everyting the learner or "player" would do or need to know. Sort of like Mavis Beacon Teaches Typing...

That's so old...

What this article is about is an adaptive neural net processor which actually learns from it's past experiences and gets better at it's job without further human intervention.

The project's ultimate goal is to create a small, low power-usage computer that analyzes complex information from multiple senses at once, but also dynamically rewires itself as it interacts with its environment.

...cognitive computers are expected to learn through experiences, find correlations, create hypotheses, and remember - and learn from - the outcomes, mimicking the work of the brain.


That's a very, very big difference in functionality and adaptability.
nicknick
1 / 5 (1) Aug 19, 2011
I wonder whether the dynamics of the plasticity of the brain are well understood yet. I know a mouse brain was cut into thin slices to make a 3D computer model of it in the BlueBrain project. Yet, you can cut a brain only once. To understand the dynamics, one should be able to observe the adaptations in connections and synapses of the brain in-vivo at different and many time intervals. Without good knowledge on the dynamics, I guess these circuits will not perform much better than the artificial neural networks we already know.

More news stories

Net neutrality balancing act

Researchers in Italy, writing in the International Journal of Technology, Policy and Management have demonstrated that net neutrality benefits content creator and consumers without compromising provider innovation nor pr ...

Cosmologists weigh cosmic filaments and voids

(Phys.org) —Cosmologists have established that much of the stuff of the universe is made of dark matter, a mysterious, invisible substance that can't be directly detected but which exerts a gravitational ...

Bionic ankle 'emulates nature'

These days, Hugh Herr, an associate professor of media arts and sciences at MIT, gets about 100 emails daily from people across the world interested in his bionic limbs.