A Brainy New Chip Could Make Computers More Like Humans

By Valerie Ross | August 19, 2011 12:50 pm

One of IBM’s prototype cognitive computing chips

What’s the News: Researchers at IBM have developed a new “cognitive computing” microchip inspired by the brain’s computational tricks. These new chips, the researchers say, could make processors that are more powerful and more efficient than today’s computers—and better at the flexible learning and responses that are a struggle for current AI systems but a breeze for the human brain.

How the Heck:

  • IBM has made two prototypes of the new chip, which it calls a “neurosynaptic core.” Both are built on a standard semiconductor platform with 256 “neurons,” the chip’s computational components. RAM units on the chip act as synapses; one of the chips has 262,144 synapses, while the other has 65,536.
  • These networks take after the brain in two key ways, says Dharmendra Modha, the project leader at IBM. The hardware for memory and computation are quite close together (as they are in the brain, where neurons are responsible for both) and the connections between them form, strengthen, and weaken based on learning and experience, just like synapses between neurons.
  • Other than that, the researchers haven’t yet shared many specifics on how the chips work or what they can do. At a conference next month, the team plans to present studies that show these “neural cores” not only use little power, but can play Pong, recognize pictures, and steer a car along a simple track.

What’s the Context:

The Future Holds:

  • In the long term, IBM wants to build a system that has 10 billion neurons and 100 trillion synapses (as many synapses and a tenth as many neurons as the human brain), uses just one kilowatt of power, and can fit in a shoebox.
  • Ultimately, Modha told Popular Science, cognitive computers would be able to combine lots of inputs and make sense of them, the way the human brain does: taking into account the firmness, color, and odor of a piece of produce, say, to tell whether it’s ripe or rotten.
  • Brain-inspired computers would be a complement to, rather than a replacement for, today’s systems, Modha told Wired.com:

Today’s computers can carry out fast calculations. They’re left-brain computers, and are ill-suited for right-brain computation, like recognizing danger, the faces of friends and so on, that our brains do so effortlessly.

The analogy I like to use: You wouldn’t drive a car without half a brain, yet we have been using only one type of computer. It’s like we’re adding another member to the family.

Image courtesy of IBM Research Zurich

CATEGORIZED UNDER: Mind & Brain, Technology, Top Posts
  • Chris

    And skynet begins

  • http://www.thealders.net Doug

    Damn you beat me to it Chris :) – Aside from that I wonder why one chip has only 25% of the RAM the other chip has – why the asymmetry?

  • http://N/a Mark

    If IBM told you they would have to kill you.

  • Cory

    @ Doug: The larger synapse unit is a one-time programmable neural pattern. Kind of like a person who could only learn one skill. The 60k synapse unit can adaptively learn a number of skills, reinforcing or degrading neural pathways much like the human memory.

  • http://www.thealders.net Doug

    @Cory – thanks – that helps

  • Rich

    If we can live long enough we’ll be able to reboot our brains and add more ram. Might curtail the affects of Alzheimers. You never know what’s coming next.

  • Daniel J. Andrews

    Never mind chips that make computers more like humans…I want a chip that makes me more like computers–access to large amounts of memory, increased processing speed, performing complex calculations in my head.

  • http://www.nicky510.com Crow
  • Clarence Knight

    I’m afraid this new type of computation becoming as illogical as most humans are. Will this new computer become a democrat,republican, catholic,Protestant ,atheist or Muslim or… ?
    Can this computer hold contradictory information without crashing?
    I understand a mind without a filter goes crazy.

  • lily

    the only missong part is emotion 😀

  • starsky51

    ANNs can simulate around 100,000 neurons on a standard desktop pc. These chips have got a lot of catching up to do. That said, ANN software can only take us so far. These chips are clearly the future of neural networks.

  • Cmdr. Awesome

    @14 Actually, neural nets learn in approximately the same way as we do…just with more math. They’re presented with input data, run it through their internal “neurons” and produce an ouput response. If they’re wrong, they go back and change their internal structure a bit and try again on a new piece of data.

    It’s not precisely the same, but they do make mistakes and they do learn from them.

  • Brian Too

    I was pleased that Modha suggested that “Brain-inspired computers would be a complement to, rather than a replacement for, today’s systems…”.

    Chip architectures are surprisingly conservative, in terms of their commercial relevance. Lots of new designs have been attempted and quite a few have come to market. However the experience of the old chip wars (CISC versus RISC, etc.) suggests that there is very substantial resistance in the market when chip architecture changes requires software changes. Something as simple as a recompile with separate binaries for different chips can cause software distribution issues.

    It’s not that the market hates change. The market just wants the interface to the chip to remain highly stable, with some optional additional features and functions being added from time to time.

    In order for the market to accept an entirely new architecture there needs to be some really compelling advantage. And history suggests, performance advantages in the range of 5-50% isn’t enough.

  • Rick

    India? Isn’t that a country overflowing with world class universities? Don’t virtually all the Nobel Prizes go there? And is IBM, a company really focused on selling software and business services, really a hotbed of cutting edge research? Oh, yeah, Watson. I’ll take Jap anUS relations for $200, Alex.

  • Praedor

    OK, let’s get this thing developed lickety split. This chip could do wonders to improve computer game AI! No more retarded or moronic game character AIs that has them stuck against a wall, following a predictable path, etc. It’s what it can do for gaming that matters, not boring accounting, data analysis, etc. That’s ho-hum been there done that stuff. Get it into gaming where the fun is.

  • http://www.johnz.realstew.com John M. Tax

    Yes, let’s teach computers to kill! Why not elevate the human mind out of the gutter so that we can stay one step ahead of it all.
    However, I do believe computers will overtake us. They are already creating parts of themselves. Once they start designing themselves we are left in the dust – just another cow in the pasture.

  • http://WISH4Humanity.org Al Irvine

    I think these new chips are exciting and will lead to new and important capabilities. But suggesting that they will result in computers behaving like human brains, is like believing that some reconstructive surgery on the vocal tract of chimpanzees would allow them to speak English.

    I was building tools for AI research and teaching AI programming techniques in the early
    1960’s. AI researchers have continued to vastly underestimate the complexity of the human brain and the difficulty of producing “intelligent” behavior via electronics. At that time one of the 5 or 10 leading AI researchers was traveling around the country, giving a talk in which he said that if someone would just fund a computer (1960’s speed cpu) with at least 8 megabytes of RAM, he could solve all of the problems of AI!

    What we have learned about the brain in those 50 years is that the more we learn about it, the more there is that we don’t know or understand, e.g. microtubules. I think that there is very little doubt that the brain is the most complex, most dynamic structure we have found. The notion of “downloading” a brain sets a new record for farfetched “predictions” of the AI community.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!


80beats is DISCOVER's news aggregator, weaving together the choicest tidbits from the best articles covering the day's most compelling topics.

See More

Collapse bottom bar