Slashdot Mirror
Artificial Brain '10 Years Away'
SpuriousLogic writes "A detailed, functional artificial human brain can be built within the next 10 years, a leading scientist has claimed.
Henry Markram, director of the Blue Brain Project, has already built elements of a rat brain.
He told the TED global conference in Oxford that a synthetic human brain would be of particular use finding treatments for mental illnesses.
Around two billion people are thought to suffer some kind of brain impairment, he said.
'It is not impossible to build a human brain and we can do it in 10 years,' he said."
So now we can feed them to the future invasion of zombies? That way we can all co-exists.
Maybe we can build the *equivalent* of a human brain (number of neural connections in software, silicon or combination), but we don't even know how the thing functionally works as it is. How are we going to model it?
When can I put my ghost in a shell?
It is some supercomputer software to simulate a brain. Still cool!
I'd be pretty concerned about the ethics of experimenting on an artficial brain complex enough to reasonably simulate a human one. "Human rights" aren't terribly well grounded, theoretically; but to the degree that they are, mental complexity seems to be a vital factor(given that we don't generally execute retarded people, it isn't the only one, but it is a big one). Being made of meat isn't obviously a salient factor, nor is being born to human parents.
An artificial brain of that complexity would be, in effect, a moral person. If you are willing to experiment on one, you might as well just use hobos and orphans and not have to wait a decade for fancy computers(though a simulation would have the huge advantage of read system state out of memory, no mucking around with FMRIs and stuff).
I've been listening "in 10 years we'll have X awesome technology", but time come and go and nothing has changed, so, i'll be expecting this artificial brain so i could drive my flying car(you know, that 3D driving thingie) to arrive at the entrance of the spacial elevator so i could bang some lunar chicks.
Btw 10 years and i still have some bad english
Slashdot ya no es que lo era!
Then you won't have to listen to the cliche that an artificial brain will always be 10 years away. No one would use eleven years in a cliche.
In 10 years we will have artificial brain, in 50 we will have fusion. In 20 we will have true AI and cyborg. And in 5 years the date estimate for the 3 above will probably not have changed by much (I say probably as we could do leap and bound forward, but at the moment I don't see that as probable).
C. Sagan : A demon haunted world:
http://www.amazon.com/gp/product/0345409469/
visit randi.org
I'm still waiting on the scores of cancer cures that have been promised over the past decade. Talk is cheap.
A Magic the Gathering Article and Forum Aggregator
"Around two billion people are thought to suffer some kind of brain impairment, he said."
Only two billion? Sounds kind of low. My estimate is more in the neighborhood of 6-7 billion.
Has TED always been about giving nutjobs a platform for performance art?
Just what do they mean by a model of the brain? I really don't think they mean anything that would actually think.
Especially if you believe the few numbers given. If it takes a laptop's computing power to completely model a single neuron then there won't be enough computing power on the planet in ten years to model an entire human brain. There aren't even enough IPv4 addresses for that. We would be talking a cluster that needs IPv6 to talk between it's nodes.
And that wouldn't account for the computing needed to simulate the I/O signals to make a simulated brain able to do anything useful.
Democrat delenda est
Maybe not. Taking the claim at face value, then we'll never be quite dead: there will be always a copy of our brain somewhere ready to be loaded into a VM by some system admin.
If it's our system admin doing the backup and restore then I don't like our chances.
I intend to live forever, or die trying. - Groucho Marx
Sinners go to /dev/null.
Put it here instead.
So, I have all these 10000x10000 TIFFs I just took of a real brain. Now what?
Guess what I mean is, the brain is not the same from a minute to the next. It modifies itself constantly. We may be able to copy the parts (although I'm pretty sure we're more than 10 years away from that) but until we can make it "run", all we have is a stopped engine. What good would that do?
Unless what we want is a brain _model_, which is what I think is meant by the article.
It probably is within reach to build a hardware equivalent of a human brain. We don't know how to architect it, but building enough custom ICs and interconnecting them is probably within reach. The right architecture for simulating neurons probably involves some huge number of fast processors with limited memory, like a graphics board.
I'm encouraged that this guy is trying to model a mouse brain. About twenty years ago, I was at a seminar by Rod Brooks. He was talking about trying to jump from insect-level AI, where he'd made some progress, to human-level AI. I asked him why he was trying to make such a big jump; a mouse brain might be within reach. He said "Because I don't want to go down in history as the person who created the world's greatest robot mouse". So instead, Brooks did Cog, a stationary robot with head and arms which tries to fake acting human and didn't really lead anywhere. Taking a smaller step might work better.
Reaching for mouse-level AI is promising. Mice and humans have about 85% DNA commonality. All the mammals seem to have have roughly similar brain components, although the size ratios of the different sections vary widely. Humans have about 1000x the brain mass of a mouse. So if we can get a solid simulation of a mouse brain, it may be mostly a scaleup from there.
The classic mistake in AI is that someone comes up with a reasonable idea, and then thinks they're one step from human-level AI. That's approaching the problem as if it were easy. Fifty years in, we can now conclude it is hard. So taking smaller bites is indicated.
When we build an artificial brain, it will be rack-mounted in 19 inch racks.
The transplant thing has been observed, but so far I think it's only anecdotal evidence (maybe a bunch of people made stuff up, but so far I'll accept the reports on face value). Not aware of big research going on about it.
But I won't be surprised if scientists finally find out that your organs (or transplanted organs) can influence what sort of foods/drinks you'd want to consume[1], or even who you want to mate with. It does make some sense from an evolutionary advantage point of view.
[1] Like fried chicken and beer: http://linkinghub.elsevier.com/retrieve/pii/S1096219000000135
And if your entire immune system can change after a liver transplant, it means you're not just getting a liver - it's not quite so "neat and clean" as that.
http://www.dailytelegraph.com.au/news/teen-changes-immune-system/story-e6frf00r-1111115390103
So if the donor's stem cells manage to leak out and help form neurons in the recipient's brain or "stomach brain"[2], why shouldn't there be changes?
[2] The Enteric Nervous System:
http://www.psychologytoday.com/articles/199905/our-second-brain-the-stomach
http://en.wikipedia.org/wiki/Enteric_nervous_system
Who is the boss? From the point of view of the ENS, the "central nervous system" (aka brain/CNS) might just be a means to keeping the ENS satisfied.
ENS to CNS: "Hey CNS go eat a double cheese burger!".
CNS: "Hmm, I feel like eating a double cheese burger, lets do a lot of complicated stuff like driving, walking etc so that I can eat that".
Of course the CNS could say, "Must resist, have to stick to diet".
I mean, fusion power has been 10 years away for the last 40some years...
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
Depends on what you mean by "functionally equivalent". A neural net is a simple self-modifying learning machine, and any detailed simulation of a network of actual neurons like the one TFA describes would certainly qualify.
Why would anyone engrave "Elbereth"?
Personally, I have a problem killing animals too, I've never understood the generally accepted insanity that includes both the fluffybunnywunny and the rabbit pie.
I guess it's somewhere around empathy. If we can emotionally relate to the intelligence it's a horrendous crime, if we can picture the deceased as an aggressor, or sufficiently different to ourselves then it's somewhere between nothing and a victory.
Current electronics do nothing to stimulate a feeling of empathy, they're tools, extensions of ourselves. Once a machine brain can talk and reason with you, you might just feel different ?
You have to see it from an evolutionist point of view. The reason why we feel pain, is because it keeps us away from dangerous things. The reason why we don't mind killing pigs is because we have to eat something. The reason why we don't kill each other is because that wouldn't be very healthy for the survival of the race. The reason there is love is because it makes producing babies easier. And so on, we are what we are because it is good for survival and most of our core morals build up on that.
The fun part of course is that little of those morals still work when it comes to computers. Death for example becomes kind of a non-issue when you can copy, suspend and resume a program. Death on the other side is a big deal with biologic things, because you can't copy them. The death of a biologic thing is pretty final, the death of a computer program is not.
I don't really doubt that we one day will be able to build a computer capable of human-like intelligence, but when we do that, our moral system will have a really hard time to keep up with reality, as its not build around logic, but for most part just on our survival instincts.
How do you define one's psyche and how is "mental health" or "mental illness" defined, and on what set of values?
Say I'm a chronic masturbator (to be in tune with the slashdot mentality) and it's considered "defective behavior" even though my body rewards me to do continue that habit.
So, he would build a synthetic copy of my brain, emulate my current state and that's it.
Now, my brain is in constant evolution, I have eroding neurons, I learn new things making new neuron-paths, which his machine wouldn't be able to the way I imagine it.
Would he allow the brain to rewrite and rewire itself? And if so, how? Are these processes well understood enough?
If they would be understood, and able to emulate, will they write "virtual medication" to influence the virtual brain to test side-effects or the propagation of a certain chemical interacting with the brain?
If the last is possible, will we end up with sentien beings who are stuck in the same state for an eternity? Wouldn't that be sortof agonizing?
I think we can keep recursing like this until someone returns 1
members of Congress can wait that long to get one.
What most neuroscience appears to be missing is that the brain isn't an electrical system, but an electro-chemical system. To my knowledge, no one has done anything to simulate how the chemical interactions work with the signal passing and processing aspects of neurology. I think it is quite apparent that there are a great many connections between the chemical balance of the human body and how well things are working in various parts of the human body work. We already have some clues in observing how stuff like lithium helps to dampen activities in the brain preventing or suppressing many results of "mental disease." So if chemical influence can have such a profound affect, I find it is more than reasonable that chemical influence can also be a profound cause.
It would appear that scientists are trying to "memory map" the brain as a computer which is simply the wrong approach I believe. Sure there will be some improvement in understanding of how some aspects of things work, but I think they will quickly reach a plateau with this approach.
Go re-read Neuromancer to see how all this turns out. Every time you turn the damn artificial brain on it's the same deadpan backseat driver.
It was disturbing to think of the Flatline as a construct, a hardwired ROM cassette replicating a dead man's skills, obsessions, kneejerk responses. ...
He slotted some ice, connected the construct, and jacked in.
It was exactly the sensation of someone reading over his shoulder.
He coughed. "Dix? McCoy? That you man?" His throat was tight.
"Hey, bro," said a directionless voice.
"It's Case, man. Remember?"
"Miami, joeboy, quick study."
"What's the last thing you remember before I spoke to you, Dix?"
"Nothin'."
"Hang on." He disconnected the construct. The presence was gone. He reconnected it. "Dix? Who am I?"
"You got me hung, Jack. Who the fuck are you?"
=S
I wouldn't say Markram's opinion is in the majority. I know the field well, and I think that either he's sitting on a lot of super-ultra-exciting results that he has mysteriously not presented at the last conferences his team went to, or, more likely, is being hopelessly optimistic (or is confusing "years" and "centuries", or something).
Hi-Resolution MRI. Just scan someones real brain and then load it onto the computer. We don't even need to know how a 'real' brain works.
There's a hard limit on MRI resolution, based on the rate at which water diffuses through brain tissue. That limit is around 5 microns. There are some tricks that might let us do better, but they tend to involve techniques that aren't compatible with live subjects (think cryogenics and antifreeze).
5 microns is enough to resolve some neurons, but not the axons and dendrites that connect them. And even if you could resolve the physical structure, function depends on chemical and electrical characteristics that don't show up in MRI at all. fMRI gives a very coarse representation of activity, at the cost of vastly reduced spatial resolution.