Raduprv is going to like this one

There's no way in hell such a limited machine can simulate the human brain. Not even a cat brain.

Quote:

Original post by Raduprv There's no way in hell such a limited machine can simulate the human brain. Not even a cat brain.

Indeed. It says in the article they are going to attempt simulating a group of about 70000 neurons. Their ultimate goal, somewhere-in-the-future (TM), is to simulate the human brain.

[edit]Darn, I was beaten to the sentence "Indeed."[/edit]

Indeed. I'm guessing that when they say that it'll simulate the brain on a molecular level, they're meaning something almost entirely different than what it sounds like they're saying.

But I don't know if it's even possible yet to simulate in near-realtime the whole brain using an abstract neural net-like structure, assuming that such a structure is even detailed enough to get close to emulating the functionality of the brain.

Yeah, I figure they're trying to get attention and hype and money for a more basic and doable project. Either that or they're crazy.

Do we even know how a neuron works, let alone how they're all connected together?

Quote:

Original post by Ravuya Do we even know how a neuron works, let alone how they're all connected together?

I'm still glad to see some people are actually putting an effort into this. This kind of research might actuall provide some insight into these questions.

There's a difference between putting effort into a building block (i.e. "what the hell does a neuron do, and how does it do it?") and ignoring that building block and going on to create a simple and elegant solution to a moving problem (i.e. "let's go make a brain in our computers!").

It's the same thing with our MMORPG maker crap; I admire the initiative but I don't know if we're really to the point where we can actually accomplish this with our knowledge of biology.

Quote:

Original post by Ravuya Do we even know how a neuron works, let alone how they're all connected together?

maybe they are planning to find out?

for the record: they are not trying to simulate a complete brain or create consiousness. just 10k neurons from some specific area of high interest, because of its high recurrence.

I think they are going to find out alot on how computers work, specifically how they behave when told to simulate 70,000 neurons. [grin]

I am skeptical of such approaches. But at the end it will probably give us some useful approximations that are valid for certain limited cases.

Well, they COULD simulate a human brain in full on that computer. Hell given enough extra memory, you could do it on a C64. Just not in RealTime.

But yeah, I looked at it there, and I expect they're testing theories on a small scale. "If we zap THIS part here in a REAL brain, then if we're right, we should get similar results at the ends in both the real, and the simulated one"


I'm more impressed with the rat brain flightly the fighter jet Simulator. Evolution doesn't need to give em bigger teeth or claws,... humans are just going to give em nukes!

Why don't they try to simulate an ant? An ant, afaik, has about 250k neurons, so they could simulate 2.5% of it with a computer, rather than 0.000001% for a human.

Quote:

Original post by Raduprv Why don't they try to simulate an ant? An ant, afaik, has about 250k neurons, so they could simulate 2.5% of it with a computer, rather than 0.000001% for a human.

I think the title of the article is slightly misleading. Simulating the human brain is simply the-ultimate-goal (TM). About the ants, the articles I read on the net mentioned a total of about 20000 neurons in all of their nervous system.

besides, part of an Ant (ANT being my initials, I don't like the idea of people playing with an Ant brain,...) brain, or part of a human brain, does it really matter?

I've never really done more Bio than I had to in highschool, (The smells make me sick and give me headaches, not to mention I don't much care for real blood, but I'll gladly play with the fake stuff) but as far as the single cells in our brains, is there much of a difference between a human, and an ant if you look at it cell by cell?


Something I've just thought of, I wonder what sort of fraction humans have as far as the total brain power of all humans compared to the total brain power of all lifeforms less than in inch or two in size?

Quote:

Original post by TalrothI've never really done more Bio than I had to in highschool, (The smells make me sick and give me headaches, not to mention I don't much care for real blood, but I'll gladly play with the fake stuff) but as far as the single cells in our brains, is there much of a difference between a human, and an ant if you look at it cell by cell?

There is probably some degree of divergence, since insects are an evolutionary branch that is not so close... But they are probably not that different either. Perhaps there are a few differences in term of cellular structure, but they should rely on the same mechanisms, and possibly even the same neurotransmitters. One thing is for sure, our neurons are most likely almost identical to those of all other mammals.

Quote:

Something I've just thought of, I wonder what sort of fraction humans have as far as the total brain power of all humans compared to the total brain power of all lifeforms less than in inch or two in size?

Thats a pretty pointless question. You could wonder if they total more neurons than we do to make a funny comparison... But the power of a neural network lies in interconnectivity. Insects have small brains, and they have limited means of communication among themselves. The computational power of a neural network is not the sum of the computational power of all its neurons... Because neurons pretty much have no computational power if you isolate them.

Quote:

Original post by Max_Payne Thats a pretty pointless question. You could wonder if they total more neurons than we do to make a funny comparison... But the power of a neural network lies in interconnectivity. Insects have small brains, and they have limited means of communications among themselves.

Yes, but questions like how does a human mind work are almost as pointless.
Besides, could be fun to chuck all the brains of humans in a blender, and then all the brains of bugs in another, and then see which can become the better chess player or something.

Quote:

Original post by Talroth Quote: Original post by Max_Payne Thats a pretty pointless question. You could wonder if they total more neurons than we do to make a funny comparison... But the power of a neural network lies in interconnectivity. Insects have small brains, and they have limited means of communications among themselves. Yes, but questions like how does a human mind work are almost as pointless. Besides, could be fun to chuck all the brains of humans in a blender, and then all the brains of bugs in another, and then see which can become the better chess player or something.

I don't find the question of how the human mind work pointless at all... Seriously, if we can create self-aware robots one day, then think about it, all of us geeks can have perfect robot girlfriends that are programmed to love us and play RPGs with us!

It definitely has very practical applications.

Well, when I was thinking of a real use for the research into the human brain, I was thinking more along the lines of repairs to the brain after strokes or something,...

but hey, whatever floats your boat.

I still say the giant Human vs Bug brain experiment would be a worthy goal,... at the very least we'll rid ourselves of all the bugs in the world. (Ah, for the sake of my argument, don't point out that it would also rid the world of ourselves,...)

Quote:

Original post by Ravuya Do we even know how a neuron works, let alone how they're all connected together?

Actually that was one of the interesting parts of the article.

Quote:

For over a decade Markram and his colleagues have been building a database of the neural architecture of the neocortex, the largest and most complex part of mammalian brains. Using pioneering techniques, they have studied precisely how individual neurons behave electrically and built up a set of rules for how different types of neurons connect to one another. Very thin slices of mouse brain were kept alive under a microscope and probed electrically before being stained to reveal the synaptic, or nerve, connections. “We have the largest database in the world of single neurons that have been recorded and stained,” says Markram.

Personally I'd be interested to hear more about all this neural research they've done!

Also in response to this machine not having enough computing power-

Have you ever seen evolution at work? Like maybe, have you ever seen the results that a genetic algorithm spits out to a complicated problem? They are always horribly messy and inefficient. And this is no surprise, because they are solutions that are found by a dumb luck. Imagine a blind guy trying to surgically remove a kidney. It's a haphazard and very messy process, but given enough time, the guy will eventually get the job done.

My point is, consider the possibility that the human brain is incredibly inefficient. It's a machine that was built by millions of years of dumb luck. Maybe there are millions of neurons that don't really do anything. Maybe we can build the same structures out of 100 neurons that currently have 100,000 neurons. In general, maybe we can build a machine that does the same thing as a human brain, using .1% of the computing power.

Quote:

Original post by Ravuya Do we even know how a neuron works, let alone how they're all connected together?

Yes, we know how they work. How a neuron works is part of Grade 9 biology at my school.