Let's be clear on at least one thing: they're building a simulation of the physical brain hardware. At least according to the article, there is no goal to simulate human consciousness.
If we simulate a physical brain, and then it talks to us and says, "Consciousness? Yes, I understand what you mean, I feel that too," is there a distinction to be clear about?
If you simulate a physical brain, but give it no input, what is the output? Consciousness? Look, fundamentally we've had "human brain hardware" for years now. A Turing machine is a Turing machine is a Turing machine. If you know the mathematical formula for emergent consciousness, there are a bunch of people who would be excited to talk to you.
Why assume it gets no input? (Why build it if not to interact with it?)
A grown, working brain has been shaped by its input – changing the connections, the chemical balances, and other things we don't yet understand (but someday will). So when talking about a 'simulated brain', I think an embedded history of meaningful input, and continued interaction, is implied.
(Also, creating such a brain and then depriving it input might be torture, as with extended sensory deprivation in natural humans.)
You seem to be unduly assuming emergent behavior. First, we do not know how to completely mirror a human neural network. Second, even if we did we don't know how the initial input affects emergent intelligence.
I'm not saying it's impossible, but I am saying that you're taking a rather unscientific leap of faith.
I take issue with your assumption that there is some 'clear' distinction between the goal of '"building a simulation of the physical brain hardware" and the 'goal to simulate human consciousness". A distinction, if any, between those goals is anything but clear.
Further, the people most interested in 'physical simulation' also tend to be 'materialists' who think a suitably precise simulation can reproduce everything interesting about a human mind, including 'consciousness'. Without any personal 'leap of faith' as to whether that's true or false, this is an interesting and deeply murky question.
Instincts, growing up, human experience. It's the software that makes us (in this case some preconfigured parameters and training data), the hardware is going to be dumb
What do "software" and "hardware" even mean in the context of the brain? All of what you call "software" is just the result of combinations of neuron patterns and chemicals, all of which could be included in a simulation.
> A large portion of our instincts are hard-wired.
Not accurate, or at least misleading. There's a lot of plasticity in the brain even at the very basic levels. Raising a kitten in an atypical environment (e.g., without edges or contrast), for instance, yields profound functional differences from retinal ganglion cells over V1/V2 up to the dorsal and ventral streams of the visual system.
We are what we are because our brains are hard-wired to adapt to typical environments. Full isolation would cause substantial aberrations -- that's exactly what you observe in real-world cases.
Your consciousness is your brain hardware. The software exists in the neuronal connections and weights (both of which change over the course of your lifetime), and if these are taken from a real brain (presumably the brain that's been sliced and scanned) then the resulting simulation will be conscious. Specifically the consciousness of the dead person whose brain was used as the model for the simulation.
That is unless you are not a materialist, i.e., you believe that the mind is the product of some non-physical plane that somehow interacts with the physical world through the brain...
> That is unless you are not a materialist, i.e., you believe that the mind is the product of some non-physical plane that somehow interacts with the physical world through the brain...
Also note, you don't have to be a dualist to think that consciousness is not a property of a brain model. You could be a materialist and think that consciousness can be given to a machine, but still think it's separate from the turing-machine model of computation. That is, hold the view that consciousness is not a property that can emerge from a computation process; but rather some other different kind of process that's not yet known.
You seem to be implying that if a philosophical position has a name, it is not stupid. I'm sorry to be the one to break this to you but that's far from being so.
"That is unless you are not a materialist, i.e., you believe that the mind is the product of some non-physical plane that somehow interacts with the physical world through the brain..."
Is there any good arguments in favor of materialism? It's always struck me as a completely ridiculous position.
To me, the best argument would be that everything else that we have looked at in the universe appears to be completely materialistic. Do you know some reason why the brain might be different?
This subject is just too philosophical for a HN comment, but I think there are at least a few problems with this argument.
First of all, almost by definition, we can't really look at anything that's not materialistic; it's just a limitation.
For instance, there's no way to prove or disprove the existence of awareness in something or someone. If a robot walks like humans, looks like humans, is it a human? Does it experience anything? Do insects have consciousness? Do animals have it? Does anyone besides me have it? Who says even I have it? How can we test for it? Can we even define it? Who says that my computer mouse doesn't have it?
Second, I believe that we only know a tiny fraction of what's out there, or of what's there to know, so it's hard to say that there's nothing beyond matter. And who even says that everything we looked at is materialistic? Is time materialistic? Can it be made from something else?
Third, there's nothing materialistic about awareness. It's not as though we can say "it's like a computer program, just more complicated". We can't say that emotions are like variables that control the behavior of a program.
Do you know any reason why awareness might be materialistic?
Awareness is materialistic because it affects the physical world. How does it affect the physical world? You thought about it, those thoughts were translated into nervous system impulses which caused you to press keys on your computer, which was translated to electrical impulses, etc.
So far, we've found exactly zero processes which we can observe (i.e., which affect the physical world) that can't be explained by physics.
Yes, there are some processes which we can observe which we can't YET explain very well. But when I am ignorant about a phenomenon, that is a fact about my state of mind, not about the phenomenon.
It seems likely that consciousness / awareness is just another physical phenomenon that we haven't explained yet. Here are some other examples of "mysterious phenomena" which it took a while to explain: fire, why the sun rises and sets, the discrepancy of Mercury's orbit (explained by relativity), discrete energy levels (explained by quantum physics).
Now, what makes you think physics will never explain consciousness?
I think I answered your first and third. To answer your second question, both the theory of relativity and quantum physics dissolve time (it's just a variable), as I understand them.
You're talking about different things. Awareness, Will, machine-like-decision-making.
Decision-making can be completely robotic. Will, or Free Will, and whether or not it's compatible with materialism, is another controversial problem. Awareness, is what you feel subjectively. I can build a robot that senses colors, and when it sees the color red, it screams "OMG! RED!!!". This robot however has no awareness what so ever. What you described is exactly the kind of thing that might happen with robots, but it has no relationship to the awareness/consciousness aspect of the problem.
> Here are some other examples of "mysterious phenomena" which it took a while to explain (...)
This is a weak argument, and really it's irrelevant to the mind-body problem. All things you described are physical states to begin with.
A more approximation of the mind-body problem might be some one saying something like this: fire is not the same sort of thing that water is. I can hold water in my hand, but can I hold fire in my hand? I can store water for later use; can I store fire for later use? Now this sort of argument could be right or wrong, but it seems like a plausible thing to say. It's not saying fire is some kind of a super-natural miracle.
I think we both agree that we don't know what awareness is; the difference is I'm saying I know what it's not: and I know it will not arise from a computer simulating a brain.
Overall I think consciousness is an interesting topic that's worth reading about. I don't mind someone being a materialist, but it doesn't seem like you appreciate the difficulty of the problem.
It's theoretically possible that one day we'll know more about awareness, and I don't know what that knowledge would be like, but I'm pretty sure it's not something that arises from running some sort of a computer program. And I can also say with about 85% certainty that it's not reducible to any of the physics we know about today.
It's also theoretically possible that we'll never know what awareness is. It might seem outrageous and somewhat a "forbidden" idea to speak, but it's a possibility. Either due to time restraints (the universe dying before we get there) or conceptual restraints (as in, maybe our minds are just not smart enough to figure it out). Unlike most other things in nature, I don't know of anyone who even came up with a hypothesis of what awareness might be. For instance, atoms where theoretically discussed long before science proved they existed, and many different models were advanced to explain and account for planetary motions (many of them were wrong, but you get the idea). But awareness, doesn't even have any model that might account for it; not even a wrong one. That's a strong indication that it might be something outside our comprehension; not much unlike the nature of space and time.
No, consciousness is the sum of the brain hardware, brain software (which is malleable and not hardwired), and input. If specific input is required to produce emergent conscious behavior, we may not have that result.
Non-materialism isn't as far-fetched as you make it out to be. In fact, no materialist account has managed to give a plausible explanation of how subjectivity and qualia emerge from physical substrates. Not so far, at least.
The short version, without references or argumentation:
Science can't prove that there aren't any supernatural requirements for a sentient mind.
However, Science has proven that most of our intuitions about our own cognition are wrong.
Since intuition is the only evidence we have that subjectivity and qualia can arise from mysterious nonphysical processes but not computational, physical processes, Occam's Razor demands that we accept the explanation which accords with proven laws of physics.
That will be an interesting outcome. What if it's not?
Can you imagine how weird it would be if we backed into proving a universal consciousness existed in a different dimension and an exactly modelled/simulated brain didn't get to participate because it wasn't given a 'soul' by that consciousness?
Or the flip side where our brain model does develop sentience as an emergent property and the spiritualists have to deal with the possibility that they are just animated meat.
Or the creepy side where you take an imprint of your current brain state, impart it on the model and suddenly the model thinks it's you?
Science fiction has explored these concepts for a while. I know that personally I'd like there to be something "special" about consciousness and/or sentience but I am rational enough to know that this isn't backed up by solid reasoning, just wishful thinking.
In a highly networked, highly automated, world the idea that consciousness could simply 'emerge' from a sufficiently constructed system is deeply disturbing.
While much of what the film "Eagle Eye" speculated on with a sentient computer connected to everything seemed pretty bogus, the ability to disrupt and render unusable a wide range of systems simulataneously by a disgruntled AI is not an experience I would anticipate as being 'pleasant.'
I really have no idea why I was downvoted for this: everything I said was factual in nature. It is an open question on whether or not we know how to create software that has the emergent property of consciousness.
Indeed. There is a huge distance to go before we can simulate things like actual thinking brains. The blue brain proejct and the proposed project use compartmental models, that essentially break down neurons to equivalent electrical circuits. These are plausible models of neurons, but there are many pieces of the biological puzzle that are elusive or highly hypothetical in these, like plasticity, brain rhythms etc.
It's not the first time someone attempts to replicate a "whole brain"; somebody used Izhikevich simplified models to simulate a cat brain (Markram had a very public spat with them) a few years ago. Those results were not very revealing - they did notice some brain waves but nothing like an indifferent catish thought.
That said, this project needs to go forward. It's the LHC of computational neuroscience. It can be completed much faster, though, all we need is computers, lots of them.
