In response to Ray Kurzweil’s comment on Quantum Computing and the brain

I thought I’d make a little note about this because quite a lot of people have been talking about this issue.

Ray Kurzweil addressed the Singularity Summit on Sunday and gave a brief summary of his opinions on some of the other preceding talks. He specifically answered criticism from others of our ability to ever model the brain using classical computing due to the presence of quantum effects in the brain. I don’t know of any supporters of this hypothesis other than Penrose and Stuart Hameroff, but maybe they are out there. He supported his viewpoint by saying that ‘The brain doesn’t factor large numbers’.

I agree with the statement that the brain is not necessarily ‘quantum computing’, but I disagree with this particular argument, because the brain does do lots of other things which quantum computers might ALSO be good at, such as pattern recognition, image processing and memory retrieval (database searching). So I think any argument as to why the brain isn’t quantum computing needs to be a bit more watertight (start by explaining decoherence for example) if you’re going to tackle this issue.

As a secondary effect, it perpetuates the myth that factoring is the only thing QCs will ever be used for. Which is sad, because a lot of smart people might have taken that impression away with them.

Advertisements

10 thoughts on “In response to Ray Kurzweil’s comment on Quantum Computing and the brain

  1. Geordie says:

    Yeah I posted a reply to a blog post on this somewhere. Here’s what I left there:

    There are AI relevant quantum algorithms.

    Many fundamental problems in AI can be naturally recast as NP-hard optimization problems–basically all discrete pattern matching problems are like this. There is a class of quantum algorithms for solving problems like this called quantum adiabatic algorithms, which are likely to be the “easiest” quantum algorithms to realize in hardware. These can (and have) been used to build learning systems. It is not a huge stretch to think that something like this could be used as a component of a powerful machine learning system.

    Also I’d just like to point out that an evolved system like the brain could use specialized heuristics that use quantum mechanics somehow. But if this is true it certainly does not imply that brains can do universal quantum computation. There could be strong selection pressure on good pattern matching but no selection pressure whatsoever on being able to factor biprimes. In an environment like this a “wetware heuristic” for pattern matching that uses QM could evolve, but the argument that we can’t factor implies no QM is used in the brain fails.

    • physicsandcake says:

      That’s quite freaky, we almost said the same thing (although your version was expressed much more eloquently).

  2. Geordie says:

    You know what’s hilarious, we had a meteor shower here a while back and we were discussing it over lunch here. I voiced the opinion that they were NOT actually meteors but the advance warships of Cardinal Syn’s invading army, which means something if you are a GWAR fan. I then watched this http://www.youtube.com/watch?v=BqZBQquAuRQ where Oderus says the same thing.

    Freaky.

  3. I tend to agree that the factoring argument is a silly one. It is certainly possible to come up with quantum analogues of neural networks, which would presumably be a relatively good model for the brain if it does indeed have quantum effects. Simply because we cannot factor efficiently doesn’t imply that our brains are not universal quantum computers (though I strongly believe this is not the case). Factoring integers is not something our brains train for, since it just does not come up often in our experience. Equally, our inability to factor does not prove that there is no classical polynomial time algorithm for factoring integers. In short, we learn nothing from our inability to solve a particular computational problem in our heads. The whole factoring business is just one giant straw man.

  4. Mike says:

    I think it is clear that the brain does not use quantum effects to carry out computational processes. Decoherence happens too fast in the brain and evolution is far too dumb to figure out how to create the states necessary for quantum computation. Evolution has done stuff like increasing the amount of proteins in the synapse, the number of dendrites/synapses and overall neuron count (plus a few other things) to increase computational capacity.

    However , the brain IS built on top of the underlying quantum mechanical nature of all the particles contained within it. Can you remove that bottom rung of the “pyramidal deck of cards” and still retain the essence of how the brain functions (i.e. consciousness)? Obviously any single quantum process probably has a vanishingly small effect on how the brain works. A single receptor protein has a larger effect, but it is still probably negligible. However, does that mean you can just neglect this lower level stuff when creating a model?

    I know the philosopher David Pearce seems to think that quantum effects are needed to sustain consciousness. Just check out some of his writing at his website. I’m not sure if I agree with some of the things he says about that, but it is interesting. He doesn’t seem to buy into the penrose-hameroff Orch-OR model, though.
    http://www.hedweb.com/confile.htm

    Kurzweil on the other hand appears to believe a much more simplified model will work just as well. In his book he argued that it isn’t even necessary to model the brain on a synapse by synapse basis. I’m skeptical of this view, though, considering how synaptic proteins have increased in number from simple to more complex organisms up the evolutionary line. Taking proteins and their phosphorylation states out of the equation may be problematic for a computer simulation.

  5. Shai Machnes says:

    I think a very good argument can be made against large-scale quantum phenomena in the brain due to the high temperature and strong decoherence effects.

    There may be quantum stuff going on at the sub-cellular / molecular biology level (such as in photosynthesis, http://www.lbl.gov/Science-Articles/Archive/PBD-quantum-secrets.html), but that does not suggest anything like quantum computation for anything, including pattern recognition.

    There’s just no way something as hot and messy as the brain can maintain a quantum state for the length of time on the scale of neural activity – milliseconds, and the size scale of brain computation – millimeters.

  6. Vince Golubic says:

    I used to think Kurzweil was sharp as a tac. He’s now showing his age and his ability to grasp the far reaching mpact that quantum effects and processes have not only in physics, but chemistry and biology.

    All things that make the brain possible.

    To make a statement saying that chemistry or classical explanations can explain the brains’ tremendous ability to process and interpret informtion is short sighted indeed and a very archaic way of looking at brain science.

    Even Lord Kelvin, from a century ago stated
    “You can understand perfectly, if you give your mind to it”

  7. Geordie says:

    Mike and Shai: I disagree with the blanket statement that that the brain is too hot & wet for QM to be used for some part of its machinations. This may end up being true but it is not an argument that rules out the possibility.

    One thing to keep in mind is that things like entanglement can be equilibrium properties of open quantum systems. Decoherence does not make a quantum system classical. It forces nature to select specific “pointer states”, ie superpositions of pointer states decay quickly. But the pointer states themselves can remain highly quantum mechanical even when the decoherence time is short and the system is in contact with environments.

    Without having a better understanding of how brains actually function there is no solid technical argument against the possibility that QM is playing a role somehow.

  8. Doru says:

    Hello,
    As a Penrose/Hameroff believer, I do think that the speculation of deriving consciousness from computational performance (quantum or not) is somehow irrelevant.
    Discrete quantum states encoded in the dendritic microtubules, do not enable the brain for computing power, but rather gives the ability to model complex space-time reality that matches pretty good the complexity of the space-time reality outside of itself.
    Computers will be always sequential in nature and never able to collapse their state for themselves (become self conscious)

  9. Shai Machnes says:

    Geordie,

    While there are certainly hot macroscopic quantum phenomena, such as the existence of semiconductors (the band structure, etc), this is a type of “pointer” state that does not allow for quantum computation.

    QC requires a controllable sustained superposition of quantum states.

    So while I cannot prove it, of course, I fail to see any way to sustain such a superposition at room temperature for 10^20+, highly entropic particles. And AFAIK, nothing even close to this (even by 10 orders of magnitude) has ever been observer.

    Shai

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