Monday 17 March 2008

Analytic and Synthetic Brains

Analytic vs Synthetic, Artificial vs Natural, Designed or Emergent....

There are two basic ways of making a system to do something. In one, you analyse what it is you're trying to achieve, then try to make something that meets one or more of those requirements, refining as you go. In the other, you have a system that meets the requirements, though you don't know the mechanism, how it does it. So you try to make something similar, and see what happens, finding out by instrumenting it what it does and how it does it.

The quest for artificial intelligence has concentrated on the first technique. Until now. The illustration to the left shows a Rat Brain neocortical column. Part of a rat's brain. An artificial one.

From SeedMagazine.com :
In the basement of a university in Lausanne, Switzerland sit four black boxes, each about the size of a refrigerator, and filled with 2,000 IBM microchips stacked in repeating rows. Together they form the processing core of a machine that can handle 22.8 trillion operations per second. It contains no moving parts and is eerily silent. When the computer is turned on, the only thing you can hear is the continuous sigh of the massive air conditioner. This is Blue Brain.

The name of the supercomputer is literal: Each of its microchips has been programmed to act just like a real neuron in a real brain. The behavior of the computer replicates, with shocking precision, the cellular events unfolding inside a mind. "This is the first model of the brain that has been built from the bottom-up," says Henry Markram, a neuroscientist at Ecole Polytechnique Fédérale de Lausanne (EPFL) and the director of the Blue Brain project. "There are lots of models out there, but this is the only one that is totally biologically accurate. We began with the most basic facts about the brain and just worked from there."
...
Terry Sejnowski, an eminent computational neuroscientist at the Salk Institute, declared that Blue Brain was "bound to fail," for the mind remained too mysterious to model. But Markram's attitude was very different. "I wanted to model the brain because we didn't understand it," he says. "The best way to figure out how something works is to try to build it from scratch."
...
The computer screen is filled with what look like digitally rendered tree branches. Schürmann zooms out so that the branches morph into a vast arbor, a canopy so dense it's practically opaque. "This," he proudly announces, "is a virtual neuron. What you're looking at are the thousands of synaptic connections it has made with other [virtual] neurons." When I look closely, I can see the faint lines where the virtual dendrites are subdivided into compartments. At any given moment, the supercomputer is modeling the chemical activity inside each of these sections so that a single simulated neuron is really the sum of 400 independent simulations. This is the level of precision required to accurately imitate just one of the 100 billion cells—each of them unique—inside the brain. When Markram talks about building a mind from the "bottom-up," these intracellular compartments are the bottom. They are the fundamental unit of the model.
...
"The simulation is getting to the point," Schürmann says, "where it gives us better results than an actual experiment. We get the same data, but with less noise and human error." The model, in other words, has exceeded its own inputs. The virtual neurons are more real than reality.
...
"We were all emotionally prepared for failure," Markram says. "But I wasn't so prepared for what actually happened."
...
This is what makes the model so impressive: It manages to simulate a real neocortical column—a functional slice of mind—by simulating the particular details of our ion channels. Like a real brain, the behavior of Blue Brain naturally emerges from its molecular parts.
"The behaviour naturally emerges from its molecular parts"... simulate the ion channels to sufficient accuracy, you get behaviour like a nerve cell. Simulate clusters of nerve cells to sufficient accuracy, and you get behaviour like part of a brain. Simulate parts of a brain with sufficient accuracy...
But there's a few practical problems right now.
In fact, the model is so successful that its biggest restrictions are now technological. "We have already shown that the model can scale up," Markram says. "What is holding us back now are the computers." The numbers speak for themselves. Markram estimates that in order to accurately simulate the trillion synapses in the human brain, you'd need to be able to process about 500 petabytes of data (peta being a million billion, or 10 to the fifteenth power). That's about 200 times more information than is stored on all of Google's servers. (Given current technology, a machine capable of such power would be the size of several football fields.) Energy consumption is another huge problem. The human brain requires about 25 watts of electricity to operate. Markram estimates that simulating the brain on a supercomputer with existing microchips would generate an annual electrical bill of about $3 billion . But if computing speeds continue to develop at their current exponential pace, and energy efficiency improves, Markram believes that he'll be able to model a complete human brain on a single machine in ten years or less.
Wow.
"If you're interested in computing," Schürmann says, "then I don't see how you can't be interested in the brain. We have so much to learn from natural selection. It's really the ultimate engineer."
Well, I am working on evolutionary computation in my PhD; and that's because I'm interested in brains. I think we'll have to evolve, to grow true Artificial Intelligence, not design and manufacture it.
Some philosophers, like Thomas Nagel, have argued that this divide between the physical facts of neuroscience and the reality of subjective experience represents an epistemological dead end. No matter how much we know about our neurons, we still won't be able to explain how a twitch of ions in the frontal cortex becomes the Technicolor cinema of consciousness.

Markram takes these criticisms seriously. Nevertheless, he believes that Blue Brain is uniquely capable of transcending the limits of "conventional neuroscience," breaking through the mind-body problem. According to Markram, the power of Blue Brain is that it can transform a metaphysical paradox into a technological problem. "There's no reason why you can't get inside Blue Brain," Markram says. "Once we can model a brain, we should be able to model what every brain makes. We should be able to experience the experiences of another mind."
...
When listening to Markram speculate, it's easy to forget that the Blue Brain simulation is still just a single circuit, confined within a silent supercomputer. The machine is not yet alive. And yet Markram can be persuasive when he talks about his future plans. His ambitions are grounded in concrete steps. Once the team is able to model a complete rat brain—that should happen in the next two years—Markram will download the simulation into a robotic rat, so that the brain has a body.
So we should see if this path is a goer within the next two years. And we may just be able to make a self-aware creature in that time frame. And that will lead to all sorts of ethical considerations. Is a Robot Rat that can pass the Turing test at the molecular level alive? I think so. It's a Rat, implemented on different hardware. It will be able to feel pain, be frightened, crave affection... And the upshot of all this, the final goal?
But the question remains: How do you know what the rat knows? How do you get inside its simulated cortex? This is where visualization becomes key. Markram wants to simulate what that brain experiences. It's a typically audacious goal, a grand attempt to get around an ancient paradox. But if he can really find a way to see the brain from the inside, to traverse our inner space, then he will have given neuroscience an unprecedented window into the invisible. He will have taken the self and turned it into something we can see.
We will be able to make a properly instrumented version of ourselves, of our minds. One that will be able to give us understanding of ourselves, how we think and why. A Mirror of the Mind.

Singularity, anyone?

But there's another, even more exciting (albeit remote) possibility. What if it doesn't work? What happens if we can simulate in enough detail all the steps leading to Mind, and we'll have cross-checked that each step of the way - what happens if no Mind results? It wouldn't be conclusive evidence, but it would strongly suggest that Mind, that Self, is not an inescapable consequence of the brain's anatomy, the way that the Mona Lisa is an inescapable consequence of a particular arrangement of pigments and canvas. It would be evidence, not proof, but evidence, that there was something more, something we've missed. Something not physical, but metaphysical, not natural, but supernatural. It would be the first evidence for a longstanding conjecture, something that would promote it to a hypothesis. The existence of the Soul.

And wouldn't that be interesting?

5 comments:

Anonymous said...

What would the researchers do if their simulated human brain realised it was trapped in a machine, became depressed and killed itself? Or would a simulated mind not have emotions? That is one of the things that will be so interesting to see.

Anonymous said...

Einstein's brain was sliced and imaged after his death, so in theory, with a big enough computer we should be able to reinstantiate Einstein.

Zoe Brain said...

Anne O'Namus : that is the biggest ethical issue. I think it likely that such an AI *would* have the normal emotions.

But fully adult humans don't come into being full-grown, they must evolve from a foetus.

The initial state will be a pre-birth infant, and then we have to come up with a way of giving such a child a way to grow up, within an environment that is totally artificial.

There are huge ethical issues, and if we're not very careful, we'll make Mengele look saintly.

Starting with finding a good dividing line to initiate the process. At what point is a foetus a person? Experimentation like this would point to a clear dividing line for ethical abortion. Or might show that there *is* no such line. We're treading on very dangerous ground here, experimentimg with people, and even children. They just don't look like other people.

One of the my first blog entries nearly 5 years ago raised ethical concerns about similar issues. As time went by, I called for consideration of those issues ASAP.

We can plausibly avoid the issue when dealing with a non-organic artificial intelligence with the same external behaviour, but we know Rats think. And the situation regarding fully inorganic artificial intelligence is not as clear-cut as it once was, given the experimentation with Cyborgs and prosthetic brain parts. There is potential for suffering on a scale undreamt-of, and for very much longer than a normal lifespan. Call it Hell on Earth. Conversely, there is the possibility that we might fully understand the nature of thought, and resolve the issues of how we should treat animals. We may even be able to augment ourselves to become, if not Gods, perhaps a little more wise as well as intelligent. Call it Heaven on Earth.

With this technique, we can't "plausibly avoid the issue" when it comes to non-biological AI.

I was thinking in 2003 that we'd have to face such things in the next 10 years. Well, time's running out. It looks like I was accurate in my estimation.

Lloyd Flack said...

You left out a possibility if it doesn't work. There need not be a soul there only needs to be processes going on that we have no handle on. Possibly something that we won't understand until there another major scientific breakthrough. Future generations need their challenges and I suspect this will be one of them.

Dave said...

actually I think the researchers are getting ahead of themselves a little. Yes, it's great that they can simulate a neuron. But researchers tend to be more aware of the obstacles in their own field and less aware of "other people's problems."

For instance, how all those neurons connect up to create something intelligent is not a solved problem. There are complex, subtle structures on many levels.

They've figured out how to make bricks, but that doesn't mean we're ready to build New York.