Slashdot Mirror
Microchip Mimics a Brain With 200,000 Neurons
Al writes "European researchers have taken a step towards replicating the functioning of the brain in silicon, creating new custom chip with the equivalent of 200,000 neurons linked up by 50 million synaptic connections. The aim of the Fast Analog Computing with Emergent Transient States (FACETS) project is to better understand how to construct massively parallel computer systems modeled on a biological brain. Unlike IBM's Blue Brain project, which involves modeling a brain in software, this approach makes it much easier to create a truly parallel computing system. The set-up also features a distributed algorithm that introduces an element of plasticity, allowing the circuit to learn and adapt. The researchers plan to connect thousands of chips to create a circuit with a billion neurons and 10^13 synapses (about a tenth of the complexity of the human brain)."
I call this microchip brain "the Pinhead" *
* small print: actual "pinheads" (microcephaly) have more brain capacity than this chip
...it starts. Looks like we're on our way towards the technological singularity.
We're all dead.
In fact, the current prototype can operate about 100,000 times faster than a real human brain. "We can simulate a day in a second," says Karlheinz.
We are SO fucking dead.
Technoli
Add a few chips and you'll soon get "I think, therefore I am."
Keep going and you'll end up with "Bite my shiny metal ass you meatbag!"
I wonder if the researchers will know when to STOP adding the together?
Learning HOW to think is more important than learning WHAT to think.
..welcome our new, silicon-brain-on-a-chip overlords!
Are YOU using the TOOL, or is the TOOL using YOU? Think about it!
The first words out of it were: "They misunderestimated me."
I've abandoned my search for truth; now I'm just looking for some useful delusions.
Oh wait. The researchers already did.
Bastards stole my thunder.
"Trolls they were, but filled with the evil will of their master: a fell race..." -- J.R.R. Tolkien on Olog-hai
Does this mean we have completed an artificial politician brain?
Violence is the last refuge of the incompetent. Polar Scope Align for iOS
Cyber zombies! "Braaains..."
This chip sounds stupid.
If the Terminator movies have taught me anything, human bones make good groundcover.
I Am My Own Worst Enemy
The more I learned about computers, the more I figured that they were more like a complex engine (data or gasoline is input, its moved around, operated on by parts, and then output as results/exhaust). Maybe that's why car analogies are so popular?
But another thing to be wary of is chemical imbalances. How many brain disorders are caused by the absence of a protein or inhibitor? The chip might take several redesigns over several years to get a solid model of a properly functioning neuron. I mean, who is going to notice a schizophrenic ant or beetle, or a rat with the mental equivalent of down's syndrome? They might spend a decade building up a brain with the complexity of a human brain only to find out that its "mentally disabled". Just look at how many people have mental issues, be it emotional, learning, or developmental issues with "properly functioning" neurons but are lacking one of a hundred chemicals that make them all work together as a whole.
I'm sure that the end result of this experimentation is not a human brain, but instead a robot that can navigate ruins like a rat (downs syndrome or not) or work together like (schizophrenic or normal) ants. I'm sure they'll eventually make a financial computer that can work like a wall street broker (employed by aig or not).
Is it sad that I am more likely to recognize you and your posts by your sig than your name or UID?
A book is a bunch of letters: A-Za-z
Having 100.000 computerized neurons is like having a "book" made of 100.000 letters. It don't mean make any sense (=It will not compute stuff, just kind of 'exist'). But could be a interesting tech bed to try to make something like, who know? maybe the brain of a worm, or the brain a snake.
I don't know a word about the topic.
-Woof woof woof!
This is nothing more than throwing more hardware at an existing problem. This has been emulated in software before, with nothing much to show for it. This will make it easier to model such things, but multiplying almost nothing by many, many times is still very little.
*My* brain mimics a brain with 200,000 neurons.
From the article: "The reason why computers seem much slower is that they are serial machines, while our brains run in parallel"
Computers are definitely faster than humans doing focused tasks, like computing a 1024-point FFT or inverting a 1000x1000 matrix.
A picture is worth exactly 1024 words.
No mention of the fact that it will become self-aware in 2 years and 25 days, or that two days later, the war on humanity will begin.
https://www.eff.org/https-everywhere
So when is Cyberdyne planning on officially launching Skynet?
I for one welcome our new stupid robotic overlords.
Toasters. Then stacked blonde religious nutjobs will penetrate our security, and it'll all be over.
Best Slashdot Co
It seems like these approaches are constrained in connection complexity by semiconductor fabrication, which would seem to severely limit the geometry to 2d. The article doesn't go into this, and it seems likely they put some effort into working around this with traditional approaches using buses and the like, but it does seem like you can't achieve the same degree of interconnection complexity on a thin 2d wafer as is seen in a typical 3d brain...
http://en.wikipedia.org/wiki/Memristor
Every time any story mentions them, their potential applications are reduced to the staggeringly, criminally mundane "could lead to faster computer memory". Standard von Neumann computer memory. A shame.
The brain is not a sequential Turing machine. Has any form of finite connectionism even been proven Turing-complete?
That (if I understand this story correctly) they here have been able to do what they have using components suited for our "traditional" computing architecture rather than the raw connectionist architecture of the brain is wonderful. It sounds like they're emulating synapses and plasticity/learning.
But the right memristors wouldn't be an emulation -- I'm not sure if they've actually made memristors with memristance profiles specifically for mimicking biological synapses, but THIS is their utility and the future. ... I'm not quite sure how this article tripped this indignant rant. I suppose I always figured I'd see this story using memristors first, but I guess that's just the next step.
We don't even know what "thought" is, except that it's a complex chemical reaction. Wake me up when we actually know what causes sentience and how it works.
Free Martian Whores!
I mean, who is going to notice a schizophrenic ant
That's the one that is walking along, waving its antennae to no one, and creeping out the other workers.
Out of modpoints but really liked a post? 1BDkF6TtmmeZ3yqXbz9yhdYVqRYnwFoXDj
OK, this chip can execute interconnected logic in hardware. It's obvious how this could also be performed in software, albeit not in a massively parallel fashion. I don't mean to imply this is a trivial task. But the fact is that brain operation depends hugely on memory. And as far as I know, we don not have even a glimmer as to how memory actually works. Maybe some vague ideas, but certainly not a comprehensive understanding. So this thing that has been developed is really, at best, more like a simple brain stem, able to execute relatively simple logic. It's probably just a overgrown industrial process controller.
It's Forbin, not Forbes.
Now you name it Skynet, and call Schwarzenegger to stop a sexy Terminator from killing some Connor guy... I' would better stop drinking so much coffe...
"O grande sentido oculto nas coisas, é não haver sentido oculto nenhum" - Fernando Pessoa
"creating new custom chip with the equivalent of 200,000 neurons linked up by 50 million synaptic connections."
"The researchers plan to connect several chips to create a circuit with a billion neurons and 10^13 synapses (about a tenth of the complexity of the human brain)."
Presumably, for very large values of "several".
I for one welcome our 200,000 neuron brain mimicking microchip overlords.
The details are not important. Like you can drive a car, and you don't need to know the color of that car to drive it. A neuron is a simple thing. It collect M signals, and generate a single output.
-Woof woof woof!
Read the great SEED article closely. The IBM Blue Brain project was trying to map the physical layout of the neocortical column, a standardized blob of nerve cells about a millimeter long and a fraction of a millimeter in diameter. If the brain is a machine made of modular parts, then the neocortical column is the starndard Lego used, over and over and over.
And let's not forget the fact that human brain isn't just a lump of neurons. It has structure, which is vital for its proper operation. It's exactly like how it's not enough to simply throw a few million transistors together to have a functional computer; they must also be connected just right. The good old Pentium demonstrated this nicely.
Forget magic. Any technology distinguishable from divine power is insufficiently advanced.
Yes you can simulate a neuron, but the point is that this chip is not doing that. What they are calling the equivalent of a neuron here is at least an order of magnitude (likely more than one) simpler than a real neuron. That is why these comparisons where they say 1/10th the brain are vastly off base. Plus the effects of the glial cells on processing is showing that they have more importance than previously thought. Since we don't really understand the brain in any great detail, all these comparisons tend to make me wince. They almost always equate very simple circuits (relatively) to neurons. It is a red flag for hype really.
If robots are ever more intelligent than us, they'll also be intelligent enough to make good decisions.
Two points to bring up.
Point the first. Intelligence does not equal good will. Don't make me Godwin this thread.
Point the second. Good decisions...for whom? Us or them? Your robots may have different notions than you have.
Weaselmancer
rediculous.
It has 200.000 neurons ? It can run for US President, having as lower standard the previous one...
"O grande sentido oculto nas coisas, é não haver sentido oculto nenhum" - Fernando Pessoa
I find that ironic given the subject matter. Or maybe it's just my font - lowercase L's and capital i's look the same in it.
Member of the 7 Digit UID Club
That would be Sarah Connor (no "O'"). The one you're thinking of was sent back in time from an alien planet to stop immortal scottish robots by cutting their heads off. Sarah O'Connor MacLeod.
"The researchers plan to connect several chips to create a circuit with a billion neurons and 10^13 synapses (about a tenth of the complexity of the human brain)."
Does anyone else think that it's a particularly bad idea to build George W. Bush out of hardware?
blog |
Stop it. Just, stop it. That is a patently untrue myth, and by posting stuff like that you are only serving to propagate it even further. Go take a look at some fMRI and PET scans, then come back and apologize. *takes a chill pill* Sorry, hit a nerve. Mindlessly repeated soundbytes like that are the bane of all that is true and valid.
I didn't read the featured article, but whenever I see "X program/system mimics brain" I always try to pipe in with my 2 cents.
Any system that considers a brain as nothing but a series of perceptron-based connections is going to fall short of the neurology of the actual brain it is trying to mimic. Ask any neurologist and they will tell you that there many other dimensions at play in the human brain. For instance, the whole system itself is sitting in a chemical bath which can change at any moment with the right mixture of hormones or other chemical changes. These changes in chemistry affect the firing and working of the neurons, axons, and synapses. Combine this with the control of external factors such as DNA, RNA, and epigenitics and things start getting exponentially complex.
I don't mean to down-play the progress we're making in this field. I just hate it when I see the "Computer system with X-sized neural network must equal a brain with X-number of neurons" mentality.
Faith is a willingness to accept something w/o complete proof and to act on it. Reason allows you to correct that faith.
Just look at how many people have mental issues, be it emotional, learning, or developmental issues with "properly functioning" neurons but are lacking one of a hundred chemicals that make them all work together as a whole.
Well... Turns out that most of our creative and historically important people have mental health issues ;)
http://en.wikipedia.org/wiki/List_of_people_with_depression
http://en.wikipedia.org/wiki/Authors_with_mental_illness
And under the depression... It cites people like Churchill and Newton.
Perhaps mental abnormalities is a motivation for them to do great things or at least stick out from the normal crowd.
You know... Like Joan of Arc ;)
"I am the king of the Romans, and am superior to rules of grammar!"
-Sigismund, Holy Roman Emperor (1368-1437)
doesn't that mean that it will be able to zap us when we try to turn it off?
We are only now beginning to figure out that there were some missing pieces in our understanding of how the brain works at it's 'bit level', for instance it is now proven that dendrites act on the neurons and have an influence on memory. There may be other cells, so far classified as 'irrelevant' that show up to have a crucial importance in the simulation of a 'real' brain... Who knows. But if they can backup my brain on that thing, I'm all for it.
Non-Linux Penguins ?
Got his number via Google from a Discover Magazine article. We can do insect brains now.
I'm sure they'll eventually make a financial computer that can work like a wall street broker (employed by aig or not).
Considering the state of the world's economy, I think they just did!
Free Martian Whores!
Its in my 20q game.
*consciously* using. That means what you decide to use not what your brain is automated to use. But given that i am not a scientist in the field of neurology can you link a laymen site that gives some information?
I do not support "The Man". I also do not support your irrational stupidity
Now if I could only recall what famous movie I saw something like this in....
I can't quite remember, I wonder why that is.....I have had a bad case of cliche dejavu.
The researchers plan to connect several chips to create a circuit with a billion neurons
So an individual chip has 200,000 neurony-type things...and so they require 5000 of them. This is an application of the word "several" I'm not familiar with.
What could possibly go wrong?
At least I hope we'll have Culture overlords... drug glands, body manipulation and uploading to a Mind, at least.
People keep thinking about it being smarter than humans and doing typical science fiction type nonsense. The real problem isn't that but instead how such a small cluster these of chips could be made into a device to crack codes, bypass security, run a botnet, and do any similar task that generally requires human input or monitoring to react to changes or to invent new strategies. Computers have been historically bad at lateral thinking in the past. Are we sure we want to give them that ability?
Think of it like a dog that moves 1000x faster than you do. You go out to get the mail and when you get back a few minutes later, it's chewed your furniture into tatters, ate all the food, dug 50 holes in the back yard, and left about a dozen piles of poop to clean up. Leave for work and come back 8 hours later...(roughly equal to a year being left alone to the dog in this case)
Obviously a computer as smart as a human causes alarms to go off and people to be wary of it. But what harm can a bunch of robots with 1/10th the IQ do?(sic for the impaired) It's the ones that fly below the radar and are seen as "benign" that are the real cause for concern.
This violates the Butlerian Jihad.
"Thou shalt not make a machine in the likeness of a human mind."
Using the Freedom of Speech while I still have it.
So I suppose, that if they scale this up, they'll end up with something about as bright as a lab rat? Rats are smart, don't get me wrong, but it hardly seems likely they'll be getting the thing to write poetry.
Terminator...hmm me thinks this is a bad idea.
(1,000,000,000 neurons)/(200,000 neurons per chip) = 5,000 chips
5,000 = (5,000/42)*42!!!!
ZOMG!!!
PS: Replace it with FOSS
welcome our silocon-brained overlords.
Sorry. Had to be done.
That's an interesting point of view that I have not thought of until now. If modeling a complete human brain, would you not also need to model the chemical imbalances? In other words. Is it possible to be self aware without a specific balance/imbalance? Is chemical interaction also modeled by this new chip?
Since most neural net modelling on smaller scales forgets that approx. 30% of brain neurons are inhbitory.
It's Sarah Connor, not O'Connor. I wonder if the Connor in the linked journal's mother is named Sarah, and if he has a brother named John?
Free Martian Whores!
Useless waste of time. organic computing is the future, this silicon crap is a joke. Why not HARVEST simple brains, jack them in, and use them as processors? Why make a metal brain when even a roach comes with a free brain to use? Hell look at the Matrix! Use human brains interconnected for a massive computing system. Felon? Death Penalty? BAH! HARVEST their brain and repurposes it as a GPU in a jar! In-game AI would be greatly improved by using REAL KILLER'S BRAINS to drive the NPCs! Nothing could go wrong unless Traci Lords throws a concert, Denzel Washington goes crazy, and the killer AI learns how to escape the virtual world and run amok looking a lot like Russel Crow while feeding off of silicates...
-=[ Who Is John Galt? ]=-
How many neurons did they emulate? A cockroach brain has been done years ago, with pretty good results. But you don't see too many intelligent cockroaches in nature. That's because *more hardware*, i.e. more neurons, are needed for intelligence. As a matter of fact, so many neurons are needed that even animals with relatively large brains fall short of human intelligence.
If there are no small brains with human-like intelligence in nature, then why should artificial small brains exhibit high levels of intelligence?
There's no animal that has both the same or larger number of neurons and same or larger brain-to-body mass ratio than humans. Considering the large survival advantage that intelligence has given to humans, it seems logical that if it were possible to have human-like intelligence with less neurons some species would have done it.
--Those things which make each and every neuron it's own miniature super-quantum-computer.
http://video.google.com/videoplay?docid=-2069501759514424839
I find it amazing that so much knowledge can be out there, but people seem incapable of connecting the dots in the ways which count.
-FL
That could be one singularity scenario, but it just takes one, right? The first one will be able to easily infer what else could have reached that point. How about many-core serial-processor chips achieving singularity through complex parallel interactions? Hybrids of the two? Really clever serial processing with a hardware RNG?
I don't care if this falls a few steps short of mimicking organic brains. I'm more happy with it appears to be a system that we can build and might find out how to get to do useful work without it having the ability to wake up on us. Later versions might be different, but the first couple of generations we shouldn't have anything to worry about. If anything, I'd be happy if this just lets us build slightly smarter microwaves, stoves, dishwashers, or washing machines. I think that those devices are already getting too wired as it is.
How could this thing be used to make our cell phone, net book, or next gen console better?
We don't want to build smart robots, we just want a maid bot that's smart enough.
You'll always have silicon irregularities, which can perhaps replicate the anomalies of DNA errors (as mentioned in another comment), as well as other imbalances. That would create regions that are in effect simulating a specific chemical imbalance. As computation switches to different regions, it is then like a specific brain area undergoing chemical changes.
I don't think the distinction between conscious and unconscious thought is clear enough that a number like 10% could be meaningful. Nor do I think you'd gain anything by bringing unconscious thought under conscious control.
Still, here's the link you wanted: link.
You want the truthiness? You can't handle the truthiness!
What governs which synapses are turned on, when? That problem only becomes more difficult when the number of "neurons" increases.
The chip might take several redesigns over several years to get a solid model of a properly functioning neuron. I mean, who is going to notice a schizophrenic ant or beetle, or a rat with the mental equivalent of down's syndrome?
Schizophrenia and Down's syndrome are HUMAN diseases; these neurons are being constructed not to build a human brain, which has a structure at least as intricate as the human BODY, but some generic processing unit.
What you are implying is a little like saying, hey we were able to join 200,000 generic cells together into a unit in the lab, but we better get all the chemicals right, or else the unit will have kidney stones.
Your point is well taken. A project like this is trying to mimic normal neuroanatomy and physiology. What happens when the brain becomes acutely or chronically pathological is much more complex. When the nervous system is in its usual healthy state, then the kind of digital architecture implemented in this project can work to a degree as a simulator. For instance, when a synapse triggers, it is in a sense an all-or-none saturated response or switch, so digital logic is useful to model it. However, neurons and their interconnections are more analog than digital, especially during conditions of disease. For instance, changes in the levels of a neurotransmitter or its receptor, or electrolyte changes which will attenuate the response to the transmitter, create variable responses, more like putting a potentiometer in the circuit, rather than an on-off switch. In fact, the responses of a neuron to its many synaptic inputs will exhibit varying degrees of time base integration, multiple input summation, and selective signal rejection, while on the output side, there can be amplification or attenuation, and oscillations or chaotic dynamics in lieu of one-shot trigger responses. This is the perfect place for large scale analog circuit modeling.
It seems like digital chips and digital programming, simulation, and control have become so entrenched (or cheap) in the mindset of designers and users, that analog gets short-changed when thinking about modern large-scale models. Designing a VLSIC with thousands of opamps, and making them addressable through thousands of addressable DAC's, and then ganging thousands of them together into your "brain computer simulator" would be a daunting and expensive chip design challenge, but ultimately far more realistic. There are some things, like the pathological states that you mentioned, that just cannot be effectively simulated on a digital chip - a software simulation yes, but not in digital-only silicon, or so it seems to me.
Did anyone else notice that their quoted 10^13 synapses can not be meaningfully compared to the number of synapses in the human brain. Each chip has lots of connections within it, but then limited connectivity to adjacent chips. While there are 50 million connections within a chip, there is little (comparatively) connectivity between chips. Some parts of the nervous system may work similarly -- e.g. sensory connections into the brain -- but connections between regions may be as complex as the connections within a region.
In other words, if they hook up 5,000 of these chips, they don't have 10% of a human brain. They have 5,000 chips each of which is .002% of a human brain and can share some data with its neighbor.
You are absolutely correct. I had a post replying to a Singularitarian (those who believe that we will be able to "upload" our selves) in the poll which covered the chemical soup modeling problem you've described as well as the I/O problem that I believe is fundamentally related. Since the other post wouldn't submit (had to re-login) I'll do some editing and put it here instead, since it is happily more topical overall.
Another thing that Singularitarians overlook is I/O. It's great that we may be able to model the structure of the human brain, but consciousness arises from and is continually affected by signals received from and sent to our sensing organs.
A human mind "trapped" in silicon would have to be somehow modified to accept an environment so utterly alien to its native one as to be likely perceived as noise, if indeed it perceived anything at all. Eyeballs work nothing like video cameras; they're much closer to specialized frequency analyzers. It would probably be less work to recreate the eyeball than to attempt to convert a video camera signal into something useful to the brain model, and the same goes for all of the other senses. A brain without a body simply isn't going to be very functional, especially when all that messy biological stuff that goes on in the chemical soup in which neurons are immersed has yet to be fully understood or modeled. Additionally, the brain's neural connections shift, shrink and grow constantly. Can a non-biological neural network do the same? (This is not a rhetorical question; I do not know the answer.)
I get the feeling that a lot of folks think we'll be able to just set up a mind, start typing questions at it, and receive answers. This view is simplistic in that it views the neural network of the brain as the only important bit of existence, when in reality we are complex patterns fully immersed and in many ways inseparable from our environment.
I used to be a Singularitarian myself (and I still enjoy fiction such as Charles Stross's Accelerando) until I read up on the fundamentals of psychology as described by William James in the late 1800s. Nearly everything in that field even today is consistent with James's discoveries in its infancy, and despite tremendous pressure to the contrary it demands that the separation of mind and body is little more than a sometimes useful fiction. Consciousness, like all sufficiently complex physical phenomena, is a dynamic process that is far too fluid for us to accurately model.
I suppose that if, as [the poster in the other thread suggests], technology will keep getting harder better faster stronger, it is conceivable that humans will eventually be able to succeed in modeling everything necessary to create a virtual environment for uploaded people to exist in, for without an environment they won't really be people (IMO they won't exhibit signs of consciousness at all). However, in addition to the hurdles I mentioned above that aren't being tackled, I have a hard time believing that technology will indeed keep getting harder better faster stronger. Maybe that's just because I'm 27 and, according to Slashdot, entering old age, but I have my reasons (see link in sig for a bunch of them). I also personally believe that following such a route is not a good idea even if possible, because we would become slaves to technology rather than using technology to better understand the mysteries of the wide reality which we confront daily.
Your brain is not a computer.
I brought up types of brain disfunctions, or, in more extreme situations, brain damage. If you're trying to create kidneys that function as a mass of simulated kidney cells, you should be worrying about simulating kidney stones and other disorders of the kidney. If you're trying to create a brain, then you should worry about brain disorders. I know dogs can have narcolepsy, would you have been happier if I had used an example that isn't just human?
Is it sad that I am more likely to recognize you and your posts by your sig than your name or UID?
Remember how in Wing Commander, AI was used to compute FTL jumps? Well, NASA could use such AI to convert imperial to metric, and vice versa, to correctly land probes on Mars! Without such technology, no other race could make accurate computations for Mars (and other solar system) landings.
If you don't know what AltaVista is (was), get off my lawn.
Data-flow programming and feed-back systems would be the closest analogy - different parts of the brain have different purposes, each area generally has a source of input and an output (receptor nerves received real-world stimuli), other nerve relay responses (spindle neurons), motor-neuron cells control muscles.
This system should allow the simulation of simple nervous systems like snails (10000 to 1 million neurons), as in many species a single neuron is known to be dedicated to a single purpose (a central pattern generator for controlling the ripples of contraction and expansion of the foot muscles). Is the purpose of a neuron determined simply by its location and connections to other neurons. If so, this system will work.
Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
http://www.gridswatch.com/index.php?option=com_content&task=view&id=1242&Itemid=14
There already is a proposed 2048 computer cluster being built. Putting 5000 of these in a massive server farm and linking them all together isn't completely beyond consideration or current technologies.(5000*200K=1 billion)
I give them 3 years, tops, until they have a cluster like this that has the computing power of a human brain. If they can increase the size to 1 million per chip in a few years(quite reasonable, IMO), then it would only need a thousand. We've done 1000 CPU clusters before. I think there's even a name for that... ;)
OMG SKYNET!!!
...the equivalent of 200,000 neurons linked up by 50 million synaptic connections
So... already more sophisticated than Steve Ballmer?
Homonyms are fun!
You're driving your car, but they're riding their bikes there.
Cloning a wetware brain is far safer. Imagine you signed up to have your brain frozen and donated to science. You say good bye to your family and take your last breaths.
you wake up and notice you are numb, and your surroundings while idyllic have very low polygon counts and no odor. A fellow shows up and comfortingly explains you are the first upload. Assuming you are currently a well adjusted individual, are you going to go nuts and kill the rest of us, your meatspace bretheren?
It seems much less likely that a machine that "remembers" its first kiss will go skynet on us, than a very well understood support vector machine that has been preloaded with psuedorandom values and then treated like a machine (no warmth, no rights, no equality, no rest from practice) until it happens upon a behavior we think we like.
refactor the law, its bloated, confusing and unmaintainable.
You're thinking about this all wrong. Just because the box looks like a computer doesn't mean we can program it like a computer. The only control one has over the neurons is to change the weighting of the signals that travel down a synapse. That's it. No "Love=1" crap available.
And the best part is after training, no one can predict what weighting affects what output in a given situation. So the only "programming" you can do is the same "programming" we give children. Can you flip a bit in your child from bad to good?
So this boils down to a problem of TEACHING, not programming. And we've pretty much proven we're only so-so at teaching (all children do not grow up to be good). So let's just HOPE we can teach this thing to be good. But we better be able to kill it if the teaching is only so-so.
Robots, just as humans, do exactly what their programming tells them to do.
Yes, but controlling the programming turns out to be impossible. So when a man goes nuts with a gun in a school, he's listening to his faulty programming. How do we know if we've installed faulty programming in a machine?
Remember that the weights of synaptic paths is ALL the programming you can do and there is no known way to determine how a machine will behave given a particular set of weightings. You can't "check" the programming any more than we could check the gunman's programming.
So expecting miscalculations is far from the stupidest idea ever.
During the training of a neural net, the weightings of the synapses are adjusted in a feedback loop until the correct output appears. Some of those weightings will go from zero to a non-zero weight (new synapses will form) and some will go from non-zero to zero (old synapses will die).
So nothing stands in the way of getting a "smart" machine by hooking up transistors in this way.
Could you tell us something about this? I haven't kept up-to-date with the field for ages.
Once long ago I spoke with a guy who was pondering whether not just the spiketrains had to be modeled, but also regional variations in neurotransmitter concentrations, which would then presumably have some kind of general effect on the neurons in that area (fatigue? lack of ATP? dunno..). It seems plausible to me that "getting tired" is an important function of a brain.
To be, or not to be: isn't that quite logical, Slashdot Beta?
Lets just hope they don't use an MS operating system as it may go straight to the luny bin in no time.