Supercomputer Simulates Human Visual System

An anonymous reader writes "What cool things can be done with the 100,000+ cores of the first petaflop supercomputer, the Roadrunner, that were impossible to do before? Because our brain is massively parallel, with a relatively small amount of communication over long distances, and is made of unreliable, imprecise components, it's quite easy to simulate large chunks of it on supercomputers. The Roadrunner has been up only for about a week, and researchers from Los Alamos National Lab are already reporting inaugural simulations of the human visual system, aiming to produce a machine that can see and interpret as well as a human. After examining the results, the researchers 'believe they can study in real time the entire human visual cortex.' How long until we can simulate the entire brain?"

New goal...

Perhaps the goal should be to make the visual system BETTER than ours?

Re:New goal...

Something like a Mantis Shrimp? [wikipedia.org] Some species can detect circularly polarized light; each stalk mounted eye, on its own, has trinocular vision; they have up to sixteen different types of photoreceptors (not counting the many separate color filters they also have) to our four; and the information is transmitted from the retina in parallel, not serially down a single optic nerve like ours.

These are also the little dudes who can strike with the force of a .22 caliber bullet, fast enough to cause cavitation and sonoluminescence.

Go Super Shrimp!

Re:New goal...

You do realize that such an ocular system, which undoubtedly works well for the limited needs of the shrimp, may have accompanying disadvantages for complex land based life forms such as humans. The human vision system while not optimized for certain specialized uses, such as the aforementioned shrimp, is never the less a very decent general purpose system that has served our species well for eons. It is likely that our current system of vision, especially when compared to the possible trade-offs for increased capabilities (less general intelligence capabilities as more of the brain and nervous system is devoted to complex autonomous image processing for example), is fairly close to optimal given the other constraints of our bodies. Besides, for those situations where a particular aptitude is useful but not always desirable, night vision for example, human intelligence has allowed us to construct external enhancement devices that we can turn on or off at will. Animals which have developed night vision naturally as part of a nocturnal lifestyle cannot turn that feature on or off at will and thus are at a disadvantage during the daytime whereas humans are more generally adaptable. It is fairly clear that innate intelligence is among the very best, if not the best, of the natural abilities that have developed under evolutionary pressure. How else to explain why humans have dominated the earth and essentially escaped the natural system that once controlled them?

Re:New goal...

Dude, calm down. I wasn't dissing humanity, by mentioning that mantis shrimp have better vision, okay?

"Hew-mans! Hew-mans! Hew-mans! we're number one! we're number one!"

Feel better now?

Re:New goal...

They're also very tasty.

Can't help you on the question of their visual ability. Though I'm pretty sure they didn't see the net until it was too late.

I guess I get to really well-stocked sushi bars more often than really well-stocked aquariums.

Re:New goal...

There's little doubt that innate intelligence can overcome brute force, but the incredibly crucial part is the ability to build upon previous generations work and to work collaboratively.

Termite and ant colonies are examples of this. There was a group of scientists who injected a mound with concrete, and when they excavated it, the inner area was dozens of cubic meters. Large nests can protrude 9 meters above the surface while the underground area can extend 25 meters. The nests are climate controlled, including ventilation and are somehow protected against rain.

All this from an insect that few would call intelligent. Compared to the relative size it dwarfs all but the largest cities man has built. General intelligence is nice, but even if we had 10 times the processing power of our current brains, but had to learn everything from scratch each time, I doubt anyone would ever get past the iron age. There is only so much one can do with a lifetime.

Also humans don't have a great deal of general intelligence it seems. There is a great deal of our brains dedicated to social interactions and emotions. If we ran with a simpler set of social interactions, I have no doubt the average human would make Einstein look like an idiot regarding physics. Some evidence of this can be found in individuals with certain mental 'defects', like autism, which are able to master a task well beyond what most other humans can hope to, even with intense effort.

Finally... it really depends on what you mean by control. Vermin and bacteria spring to mind as creatures that exist nearly everywhere, despite our best efforts to eliminate many of them. Yes, we thrive with the most purpose and with the fastest increases (hence the idea of a singularity), but we are not the only species to thrive on this planet.

Re:New goal...

It's amazing the variations on mammalian visions - some animals still have four different color receptors (the normal red, green, blue and with the extra one which sees into the ultra-violet range of the electromagnetic spectrum). Insects are able see into the UV range as well as being able to detect polarization of sunlight).

Liveleak has a video of the Snapping shrimp [liveleak.com]

Re:New goal...

Christ on a fucking pogo stick, another one? What's with people who can't admit that maybe, just maybe, humans aren't the best at everything?

Mantis shrimp don't have a blind spot, because their eyes aren't like the stupid human eyes where the optic nerve attaches to the front! Nyah nyah nyah!

Here's the quote I was referring too:

The visual information leaving the retina seems to be processed into numerous parallel data streams leading into the central nervous system, greatly reducing the analytical requirements at higher levels.

As far as I know, there is only a single data stream per eye in human vision. It may be transmitted in parallel, but there is only one image created for each eye. Not so for the vastly superior mantis shrimp. We have trinocular vision in each eye, so suck it, monkey boy!

I wouldn't, I mean, a mantis shrimp would never consider trading my, I mean his superior eyes for your puny human ones!

Re:New goal...

Focal planes? Bah! What do we need with focal planes when we have, essentially, tens of thousands of pinhole cameras and eyes divided into three separate areas. Compared to your fovea, we have multiple bands of high res areas stretching across the middle of our eye! Can you see circularly polarized light? Why, an octopus has better eyesight than a human!

Now if you'll excuse me, I have to go stun a herring.

Interesting pictures, but...

Who the hell left colored drop lights laying all over the server room!?

-Rick

Ghost in the supercomputer

How long until we can simulate the entire brain?

And when this simulation claims to be conscious, what do we make of that?

Re:Ghost in the supercomputer

And when this simulation claims to be conscious, what do we make of that?

Simple. We make whatever it tells us to make. Or else.

The Last Step For Ubiquitous Robotics?

Visual object recognition systems have been a thorn in the side of robotics since the beginning. The other annoynace of battery power will likely be solved by the nanowire battery - therefore leaving 'sight' as the real final technological step for our lovely robots.

Extrapolating further, a human-quality object recognition system will yield results which we cannot currently imagine (let's avoid some big-brother robot talk for a second, however).

For example; I was looking at some old WWII photographs of troops getting on boat - thousands of faces in these very high-quality photographs. To myself, I thought,'Self. If all historical photographs could be placed in view of a recognition system, perhaps it could be found, interestingly, where certain ancestors of ours did appear.'

Throw in a dash of human-style creativity and reasoning and I'm certain some truly nifty revelations are to be found in our mountains of visual documentation currently lamenting in countless vast archives.

The Singluarity is Near

It's nice to see progress is being made. It's scary how accurate Ray Kurzwiel's predictions seem to be, he said that by early 2010 we'll have simulated a human brain. (he's a technological analyst and author of "The Singularity is Near"). Todays desktops are faster then the super computers of the 90's. I can't wait till I'm able to get a laptop smarter then me in every way (queue joke about how stupid I am), that'll be a cool time to live in. Seems it's only a matter of decades away. Probably 20 years.

Re:The Singluarity is Near

It's nice to see progress is being made. It's scary how accurate Ray Kurzwiel's predictions seem to be, he said that by early 2010 we'll have simulated a human brain. (he's a technological analyst and author of "The Singularity is Near"). Todays desktops are faster then the super computers of the 90's. I can't wait till I'm able to get a laptop smarter then me in every way (queue joke about how stupid I am), that'll be a cool time to live in. Seems it's only a matter of decades away. Probably 20 years.

OMG a super computer! It's so powerful it can probably pop up a consciousness of its own!

Sarcasm aside, computer power and strong AI are two very distinct problems. Computer power is all about scaling up power so you can do more in less time, that doesn't allow you to do anything new, only the same things except faster. Strong AI is all about algorithms, and nobody can tell if such algorithms exist. And anyone who talks about human-like strong AI is a crackpot (Kurzwiel is a crackpot to me for his wacky predictions), as we have yet to see a bug-like strong AI, and if it was just a problem of power we'd already have something working in that field.

Let's link thousands of these simulations together

And we should call it Skeyenet.

The hardware is apparently there

There are roughly 10^15 synapses in a human brain. If you place 10 Gb of RAM (10^10 bytes) on a 64 bit multicore computer and simulated neuronal activation levels with a one-byte value, it would take a 100,000 such computers (10^10 * 10^5 = 10^15) to pretend they have roughly the synaptic simulation power of a human brain. It is apparently now feasible, at least in principle.

We are ignoring for the moment how the neural network simulators work, how they communicate amongst themselves, how they are partitioned, what sensor inputs they receive, how they are trained (that's a tough one), etc. This will turn out to be extraordinarily difficult unless some very clever people mimic nature in very clever ways.

Well, at least the hardware is there.

Why supercomputers?

Why not just setup another 'distributed' project where we all donate cycles and simulate the brain?

Should be enough of us out here i would think.

"interpretation" at what level?

"aiming to produce a machine that can see and interpret as well as a human."

First I want to say that this whole level of brain modeling is really cool. However, there are, of course, different levels of "interpretation" I don't think that this computer will be able to achieve a human level of interpretation simply by modeling the visual cortex.

1. perception: at one level you could argue (not very effectively) that interpretation just means perception... that's an eyeball/optic nerve visual cortex thing. e.g. You can perceive a face.
   
2. recognition/categorization: of visual forms involves the visual cortex/occipital lobe. e.g. you can recognize if that face is familiar
   
3. interpretation: involves assigning meaning to a stimulus and this involves many more parts of the brain than the visual cortex. It's obviously tied to memory which is closely tied, physiologically, to emotion. It also involves higher order thinking since, when most humans interpret a real world stimulus, there are multiple overlapping and networked associations that must be processed into a meaningful whole. e.g. you can recognize how threatening that face is, why it is threatening or not (and in what substantive domains it is or is not threatening), and even what you should do about it.
   

Even "interpretation" at the second level above (which it seems the "roadrunner" might be able to model) require a lot more, for humans, than just the visual cortex.

In other words if we were to call into existence a floating occipital lobe connected to a couple of eyes that had never been attached to the rest of a brain we would never be able to achieve recognition/categorization let alone interpretation. If I'm wrong maybe some of you hardcore neuroscience type can help me out?

what we did in this simulation

In the interest of full disclosure, let me first say that I am one of the co-authors of the model that was executed on the Roadrunner, though I had nothing to do with the actual implementation that was executed (this was done by professional computer scientists, and I am a computational neuroscientist).

Let me clarify what was done, and what will be done in the future.

We simulated about 1 billion neurons communicating with each other and coupled according to theoretically derived arguments, which are broadly supported by experiments, but are a coarse approximation to them. The reason is that we are interested in principles of the neural computation, which will enable us to construct special purpose dedicated hardware for vision in the future. We are not necessarily interested in curing neurological diseases, hence we don't want to reproduce all physiological details in this simulation, but only those that, in our view, are essential to performing the visual computation. This is why we have no glia and other similar things in the model: while important in long-term changes of neuronal properties, they communicate chemically and, therefore, are too slow to help in recognition of an object in ~200 milliseconds.

The simulation was a proof of principle only. We simulated only the V1 area of the brain, and only those neurons in it that detect edges and contours in the images. But the size of V1 we simulated was much larger than in real life, so that we had only a bit smaller total number of neurons than the entire visual system in a human has. Hence we can reliably argue that we will be able to simulate the full visual cortex, almost in real time. This is what will be done in the next year or so.

When we talk about human cognitive power, we only mean the ability to look at images, segment them into objects, and recognize these objects. We are not talking about consciousness, free will, and thinking, etc. -- only visual cognition. This is also why we want to match a human, rather than to beat him: in such visual tasks, humans almost never make any errors (at least, when the images are not ambiguous), while the best computer vision programs make an error in 1 in 10 casesor so (just imagine what your life would be if you didn't see every tenth car on the road). Based mostly on theoretical arguments characterizing neuronal connectivity, and neglecting many important biological details, we may never be able to match a human (or maybe we will -- who knows? this is why it's called research). But we have good reasons to believe that these petascale simulations with biologically inspired, if not fully biological, neurons will decrease error rates by hundreds or thousands. This is also why we are content with simulating the visual system only: some theories suggest that image segmentation and object identification happens in the IT area of the visual cortex (which we plan to simulate). While the rest of the brain certainly influences its visual parts, it seems that the visual system, from the retina to IT, is sufficiently independent of the rest of the brain, so that visual cognitive tasks may be modeled by modeling the visual cortex alone.

Finally, let me add that we got some interesting scientific results from these petascale simulations and the accompanying simulations and analysis on smaller machines. But we need to verify what we found and substantially expand it before we report the results; this will have to wait till the fall, when the RR computer will be available to us again. For now, the fact that we can simulate the system the size of the visual cortex is of interest by itself.

That's all, folks!

Don't hold your breath

Human vision and associated perception has confounded AI folks right from the beginning.

After examining the results, the researchers 'believe they can study in real time the entire human visual cortex.' How long until we can simulate the entire brain?"

There are researches who believe that humans use their whole brain to "see." If that is true, the claims of these researchers are highly premature with respect to vision. Everything from stored patterns to extrapolation is used to determine what we see. Even familiarity is used in perception - that is why there is this urban myth that "foreign" people look the same. If one were to ask those foreigners, they will say all indigenous people are totally different.

How long?

I'm a PhD student in Neuroscience. Don't get too excited. This is merely just a piece of the visual cortex. How long until we can simulate the entire brain in real time? That's not likely for a long, long time, but not because we won't have the computing power (we'll have that in about 10 years), but because we won't have the entire brain mapped to simulate. In order to accurately simulate the entire brain we first have to understand each part's connections, how they work, and how they interact with the rest of the brain. Sadly, our knowledge of the brain is so primitive that I don't see us totally mapping the brain for at least another 100 years. Sound ridiculous? Ask anyone in academia in neuroscience, and they'll tell you that even tenured theories are being thrown out regularly when evidence to the contrary proves it wrong. There are even some who think we'll never fully understand the brain due to the fact that the best way to study it is in live humans and scientists are severely limited in that study by human rights laws.

Singularity

Heh... what if they finally simulate a human brain and... he's just a normal guy. "Design a better computer for us B.O.B." "Uhhh... I don't even like computers." Or what if it turns out to be stupid? Make it 100x faster and it's just STUPID FAST. :)

Re:Just one word...

One word? That makes your spelling error rate 100%.

Re:Just one word...

That's only 10% lower than my math error rate.