Computers and brains get closer together: Brain-like computing components

The brain is like a computer; this is the fundamental metaphor at the heart of 1980s cognitive psychology. To an extent this was a useful way of thinking about the brain, it certainly stores and processes information just like a computer, and you can even (perhaps) draw some rough parallels between parts of the brain and computer components.

However, in at least one important respect, the brain appears to function very differently from a computer. A computers’ processing power is highly centralised in a single processor (or perhaps a dual/quad core processor – doesn’t matter – still centralised). The processor does all the computational work, and the hard disk stores all the data that the processor works on. This means that data is constantly being shuttled back and forth from the hard disk to the processor (using the RAM as an intermediary, to avoid the hard disk spinning up and down all the time) and this transfer of data is slow, inefficient and creates a bottleneck which restricts the maximum speed at which computers can run.

The brain, on the other hand works very differently. The brains computational power seems to be massively distributed. Specific functions are localised within regions of the cerebral cortex, but each area seems to have its’ own ability to a) make computations, and b) store the information it needs, at least temporarily. In other words, the computational substrate of the brain is the same as the storage substrate – neurons both store, and compute information at the same time. In theory this is much more efficient as it doesn’t require any moving around of data. Exactly how they do this is a bit of a mystery of course.

This post has been inspired by a paper just published in the journal Advanced Materials (not my usual kind of journal, but what the hell). The researchers in this paper describe a new type of electronic component which appears to mimic the dual-function of the brains neurons. They’ve used ‘phase-change’ materials to create a processor which can perform the four basic arithmetical functions (add, subtract, divide, multiply), and also stores data by the precise state of the materials’ crystallisation. This is fabulous stuff. Obviously it’s a very simple implementation at the moment, but it’s a step towards a ‘brain-like’ computing system, and the researchers say their next step is to link together a hundred or so of these chips and attempt to create a neural network for some simple tasks, like image recognition.

Future computers might look very different from current ones; they might make a modern computer based on Integrated Circuits look as clunky as the 1950s vacuum-tube monsters look to us now. Imagine opening a computer case and not seeing a system of discrete components, but instead seeing a big, homogeneous block of crystal, or a bag of goo, all of which contributes to massively parallel calculations and has a huge, distributed, storage capacity as well. Cool stuff.

Advertisements

About Matt Wall

I do brains. BRAINZZZZ.

Posted on June 28, 2011, in Cool new tech, Hardware and tagged , , , , , . Bookmark the permalink. 11 Comments.

  1. A tenet of AI/brain simulation research is this idea termed by Douglas Hofstadter to be, “liftability of intelligence” – the belief that if we understand low-level events well enough (eg. individual neurons firing), we can simulate or mimic higher levels of brain organization without worrying about what’s going on below. So do you think that creating neuron-like components are necessary to simulate the brain or to create human-like intelligence with computers?

    • Hi, thanks for the comment. You raise some interesting points. I think one of the ultimate goals of AI research is to be able to maintain a conscious entity across a range of computing substrates – in theory if the ‘mind’ is just software, then any sufficiently advanced hardware should be able to run its program – could be biological, silicon-based, quantum-computing, whatever. From that perspective a brain-like-computer would not be necessary – any general-purpose computing device could do.

      Having said that, we’re a very long way from that point at the moment, and it could be that drawing inspiration from the way that evolution has crafted efficient computing substrates (i.e. brains) might be a fruitful way to proceed, not just for AI researchers but for computer scientists in general. Building a computer which can do as much efficient parallel processing as the human brain would be a very worthwhile objective!

      • Clearly, while there are some similarities between the processing which takes place in AI applications and the general understanding that we have of how the brain processes the input from the various support systems in the human body, to suggest even “in theory” that the “mind” might be just software is so far from being credible that it casts the entire discussion into the category of speculation with a vengeance.

        I cannot understand how any reasonable person engaged in research of artificial intelligence could ignore the profound implications of their own subjective experience of their own consciousness…as a “conscious entity” without concluding that no matter what else humans may invent that “mimics” the activity of neurons and memory storage and retrieval in the human brain, could even entertain the idea that such enormously complex and richly textured experiences which result in a fully functional cognitive human person could somehow be translated or recreated in silicon or any other medium.

        Whatever degree of sophistication in computing power and capability may result from such research, a synthetic or otherwise manufactured inanimate object composed of whatever artificial substance might be produced in the future, however cleverly designed to imitate the activity of neurons and brain functions, at best, might only be astonishing in the ability to give us the sense of its being “conscious,” while failing completely in actually “being” conscious.

        Being human consists of a great deal more than “parallel processing,” and frankly, I find it sadly disappointing that some of our finest achievements and profound qualities as living beings can be reduced to “parallel processing,” by clearly intelligent and conscious humans.

        I am absolutely in favor of research that increases our knowledge and benefits humanity with results that enhance our “human” experience of being conscious, but even entertaining the notion that something manufactured in silicon might somehow be equivalent to a human being simply flies in the face of our own experience of our humanity.

        John H.

  2. Hi John,

    I agree that a lot of discussion about ‘true’ (i.e. conscious) AI is pretty speculative at present, but that’s partly what makes it so exciting. I’m very much an interested observer in that kind of field so can’t claim to be an expert at all, however, it seems self-evident to me that if we could simulate the workings of a human brain with a sufficient amount of precision, then that simulation would essentially be human, with all the attendant phenomenology associated with that, including a self-reflective consciousness.

    It’s possible that we’ll never get there – that we’ll never be able to manage an exact simulation for technical reasons. However if history teaches us anything it’s that it’s unwise to make premature judgements about what technology can and cannot do.

    I’m curious as to what you think such a simulation might be missing, that would prevent it being fully human or conscious. Surely, an exact simulation of a human brains *is* a human brain, in every important sense. Unless you think there’s something else (non-physical?) which produces consciousness, it should surely be conscious.

    Like you I’m proud of the fine achievements and profound qualities of humans – however I’m just as excited by what future artificial entities might create and produce. What’s even more exciting is that artificial entities might not have some of the limitations that humans have. A human-level of intelligence might just be the first step for such entities.

    Anyway, all very speculative – but undeniably fun.

    M.

    • Hello again, Matt,

      What seems to be missing from the research in AI simulation of brain function, aside from a realistic acknowledgement of the extreme difficulty in precisely reproducing the degree of complexity in just the structure of the brain, which is composed of naturally occurring biological matter and supported by complex biological processes which took millions of years to result in “conscious” humans, is the acknowledgement of the evolutionary progress that resulted in “conscious” modern humans in the first place.

      The earliest branches of hominids all had various versions of our modern day brains that were LESS capable cognitively, until Homo sapiens finally evolved, and even WITH a biological brain supported by the very same sensory support and comparable central nervous system as the one we possess currently, truly and fully functional “conscious” humans also took thousands of years to even BEGIN to demonstrate that they were “conscious.”

      The human brain is not simply a Hodge-podgy conglomeration of neural networks that you can slap on a motherboard and make conscious. Even with a superior technological advancement that might approach parity in complexity to the trillions of cells that make up a brain, there is also a supporting system of glial cells, blood circulation, oxygen supply system, central nervous system, etc., not to mention the influences of thousands of years of evolution on things like our emotional and psychological dispositions, as well as maturation of every cognitive capacity from untold millennia AFTER finally achieving structural and biological maturity.

      Even modern day children, fully functional in every way structurally, take years to actually become conscious, and as modern day biological constructs, they didn’t need to evolve for millions of years first. But what modern day children DO have is a rich inheritance from that evolutionary process both structurally and genetically, which provided them with the ability to EXPERIENCE consciousness, and be “self-reflective.”

      Your statement: “…if we could simulate the workings of a human brain with a sufficient amount of precision, then that simulation would essentially be human, with all the attendant phenomenology associated with that, including a self-reflective consciousness,” posits a simplicity to the existence of human experiential consciousness which totally undermines not only the complexity involved in achieving it, but reduces the “attendant phenomenology ” of experiential consciousness to a “side-effect” of brain structure.

      Consciousness is not “produced,” it is “experienced,” It is, in my opinion, achieved by virtue of existing first totally outside of physiology, and then accessed through the acquisition of an adequately equipped mental organ. It may be a fundamental property of the universe like electromagnetism or gravity, or it may be even more outside of our direct perception like the strong and weak nuclear forces, or something even more mysterious which we have not as yet understood.

      Being human and experiencing consciousness is possible due to an enormously complex biological and cognitive evolution that has granted those of us with a functional mental apparatus access to a PROFOUNDLY dynamic synergy of both form and substance which artificial means might someday “imitate” but which will never be “human.”

      But yes….undeniably fun… John H.

      • You’re absolutely right John – there must be a lot more to consciousness than just wiring up a bunch of neurons and flicking a switch. It’s likely that my view is hopelessly naive and just simulating a brain would be insufficient – however, we won’t know unless we try, and I’m positive we’d learn something useful in the process, even if it failed. Whether we’ll ever be able to simulate a brain accurately enough or not, I don’t know, but what does seem certain is that within our lifetime, the complexity of the simulations and systems that we can produce will be approaching the point where they’re similarly complex to brains. What happens then might end up being more of an empirical question! Just flick the switch and see what happens…

        There could be levels of complexity going on in our brains and bodies that we currently have no inkling of, of course. In any case, exciting times to be alive, and I honestly can’t wait for the first ‘potentially conscious’ artificial system to come about – as I said, whether it really is conscious or not, either way, we’ll definitely learn something from it.

        M.

  3. I agree with you, Matt – we won’t know what we are capable of doing unless we try and I hope it is clear that I am in FAVOR of trying! It really isn’t a matter of failing or succeeding that is at issue in this matter, since it is my contention that by attempting to simulate a brain, we will eventually discover that there is so much more to creating a functionally conscious entity than just simulating our brain structure.

    What concerns me mostly is the apparent “assumption” by many who advocate this research that nothing more will be needed other than to simulate brain structure precisely ENOUGH in order to produce a conscious entity. I would not advocate such an approach in ANY endeavor, and in my own pursuit of researching how we can come to understand consciousness better, what I see as the biggest obstacle to progress are the “assumptions” on BOTH sides–the scientific and the speculative or metaphysical camps. We cannot hope to make any significant progress in either direction by resolutely dismissing or undermining the other viewpoint.

    In my blog, I consistently try to present ideas that are supported by sound reasoning and the current scientific wisdom. I also periodically review the work of particular individuals who value the inclusion of thinking which reflects an open-minded approach to understanding consciousness, My own research has consistently reaffirmed the BENEFITS of remaining open to ideas OTHER THAN my own, and I consider conversations such as the one we have been having as enormously valuable in coming to terms with this important subject.

    Your work with fMRI technology is enormously important and critical to our understanding of brain physiology and function, and reviewing your published papers, while they are a bit daunting in their technical emphasis, demonstrates a lively and engaging intellect that reassures me there is hope for neuroscience. Your generosity in sharing your thoughts with me is greatly encouraging to me personally, and I hope you find some benefit to your own thinking from our exchanges. I also can’t wait for the eventual attempt to bring about a “potentially conscious artificial system.”

    No matter what the outcome, these are definitely exciting times to be alive….and conscious……John H.

  4. Many thanks for the kind words about my work John; though I really think you should only try reading my papers if you’re having particular trouble getting to sleep at night. ;o)

    Many thanks to you also for the dialogue – it’s exactly the kind of thoughtful and interesting interaction I was hoping for when I started this blog.

    M.

  5. Hi, I love your graphic. I need to do a pp presentation for a course and would like to know where you got it from. Thanks!

  1. Pingback: 12SB Tehnica mestecatului nu hraneste | My Wordpress

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: