Brain downloads 'possible by 2050'

[dorbie 0]

The whole brain download idea is an old sci-fi one.

Nothing demonstrates that it can actually be done or that even if a scan & simulation were attempted that it would behave like a real brain.

Read "The Emperor's New Mind" by Roger Penrose. You'll see there may be quite a bit more to simulating a brain than pulling an idea out your ass or even borrowing someone else's old idea as Ian Pearson "futurologist" and Sony shill seems to have done.

[billvon 3,120]

>In these cases if the "person" interacted in the real world...Maybe a job
>as a computer consultant...Would they have the right to vote?

I would think that you would have to start considering "partial people" under the law i.e. every mind has the right to vote ONCE, even if there are ten copies of that one mind.

>Don't know. I am not sure society would be able to handle it. It
>would depend on the level of intelegence. If they were what we would
>consider "intelligent" they would have to be integrated in some way.

I don't think that would work. Heck, we can't even decide if dolphins are intelligent or not, and we've been working with them for 50 years. A race of hominids less intelligent than dolphins but more intelligent than orangutans? Would you want them carrying guns or driving cars, for example? You'd end up with a two-class society.

[likearock 2]

I think the biggest problem is not computer speed or I/O ports but how the hell you can even represent thoughts or feelings in a digital manner. No one really has a clue on how to do that. The advances in hardware will make for better video games and more accurate weather forecasts but it will take a breakthrough of an entirely different kind to achieve a downloadable mind.

Wayne

[Ron 10]

Quote

I think the biggest problem is not computer speed or I/O ports but how the hell you can even represent thoughts or feelings in a digital manner. No one really has a clue on how to do that

We don't know how to tell what a woman is thinking now....How does that change anything?

"No free man shall ever be debarred the use of arms." -- Thomas Jefferson, Thomas Jefferson Papers, 334

[mr2mk1g 10]

Quote

ow the hell you can even represent thoughts or feelings in a digital manner

1xx horny
x01 need beer
x10 need to jump
x11 need sleep
x00 need food

did I miss anything?

[likearock 2]

Quote

Quote

I think the biggest problem is not computer speed or I/O ports but how the hell you can even represent thoughts or feelings in a digital manner. No one really has a clue on how to do that

We don't know how to tell what a woman is thinking now....How does that change anything?

Admit it Ron, that's the real reason you want this stuff to work!

Wayne

[billvon 3,120]

> how the hell you can even represent thoughts or feelings in a digital manner.

How do you represent thoughts or feelings in a few pounds of meat? The issue isn't really "it can't be done" because we do it every day. The issue is "how hard is it to copy that function in hardware." If you can prove that _you_ have feelings, then a sufficiently accurate copy of you will have similar feelings. And we are getting pretty good at simulating the physical world inside machines.

>but it will take a breakthrough of an entirely different kind to achieve a downloadable mind.

We're making them. Scientists are connecting more and more hardware to people's brains. Blind people can now see, and deaf people can now hear (to some degree) with implants, and amputees are directly connecting their nervous systems to their artificial limbs. Technologies like PET scanners are able to show brain activity directly. We now have computer based neural networks that can learn from their mistakes. It will be a while before we can just 'download' people, but the question "can that machine think?" will get harder and harder to answer as time goes on.

[murps2000 86]

Pearson said that computer consciousness would make feasible a whole new sphere of emotional machines, such as airplanes that are afraid of crashing. ***

Will they be afraid to land at Hong Kong, too?

[dorbie 0]

Quote

Pearson said that computer consciousness would make feasible a whole new sphere of emotional machines, such as airplanes that are afraid of crashing. ***

Will they be afraid to land at Hong Kong, too?

Will they be afraid to let men of Middle Eastern origin board them?

[AMax 0]

Quote

I think the biggest problem is not computer speed or I/O ports but how the hell you can even represent thoughts or feelings in a digital manner. Wayne

Representing the activity of simple neuronal networks in a digital manner is relatively straightforward. There are a few groups that already made a lot of progress in this direction. Essentially the communication between neurons in a network is a bunch of discrete signals where every input (synapse) is either silenced or gets triggered at certain frequency. However, modeling the activity of human brain will be extremely challenging because of the following reasons:

1) There are approximately 100000000000 neurons in brain.
2) Anatomically, the brain is composed of multiple regions and each region is formed by several different types of neurons.
3) The neurons form multiple synaptic contacts within the same region of brain and make projections to the other regions. The number of synaptic contacts per neuron depends on particular cell type (up to several thousands per neuron).
4) Synapses constantly form and disassemble in activity-dependent manner (not to mention developmental regulation).
5) We are still far from understanding how the damn thing works

Plus, guade already mention I/O problem. I just cannot imagine how to plug in and read simultaneously from 100 billion cells.

[likearock 2]

Quote

> how the hell you can even represent thoughts or feelings in a digital manner.

How do you represent thoughts or feelings in a few pounds of meat?

Good question! And if we knew the answer to that we'd be in a much better position to talk about downloading thoughts and real artificial intelligence. Unfortunately, we know very little compared to what we'd need to know.

Quote

The issue isn't really "it can't be done" because we do it every day. The issue is "how hard is it to copy that function in hardware." If you can prove that _you_ have feelings, then a sufficiently accurate copy of you will have similar feelings. And we are getting pretty good at simulating the physical world inside machines.

We're very good at simulating that part of the world in which we understand pretty well the rules of the game. For instance, how air pressure effects weather. Or how explosions can be represented by differential equations. We just don't know the rules yet for how the mind does its thing.

Quote

>but it will take a breakthrough of an entirely different kind to achieve a downloadable mind.

We're making them. Scientists are connecting more and more hardware to people's brains. Blind people can now see, and deaf people can now hear (to some degree) with implants, and amputees are directly connecting their nervous systems to their artificial limbs. Technologies like PET scanners are able to show brain activity directly. We now have computer based neural networks that can learn from their mistakes. It will be a while before we can just 'download' people, but the question "can that machine think?" will get harder and harder to answer as time goes on.

I think it's the same problem. Once you've cracked the secret of how the physical impulses of the brain and nervous system map to actual thoughts and feelings, you can do both. However, in spite of our great advances in understanding and mastering other parts of the anatomy and even the non-conscious parts of the nervous system, the brain itself is elusive.

Maybe, as someone said, there's just something paradoxical about trying to understand the very instrument of understanding itself.

Wayne

[billvon 3,120]

>We're very good at simulating that part of the world in which we
> understand pretty well the rules of the game. For instance, how air
> pressure effects weather.

Good example, because we don't really understand weather on a detailed basis. We don't understand what any given chunk of air will do to a developing storm, but we can guess with pretty good certainty what the storm overall will do over the next four hours. Our understanding of the brain is like that - we can predict that if you damage the visual cortex in a certain place you will lose the ability to see certain kinds of patterns, but we're not always sure how each piece contributes to the whole.

The problem isn't that we don't understand neurons, or even how they work together, because we do. The problem we are facing is that we don't know the program. It's like understanding how computers work at a very deep hardware level, and then being asked to explain how an autopilot can land a 747 in a storm. Knowing how the high speed synchronous processor bus works doesn't tell you what algorithms the autopilot is using.

One approach is to ignore the details and work on the simulation. You can take the autopilot's program and run it on a simulator, one that 'fakes' the air data and navigation inputs, and see what it does. If your simulator is good enough, then the autopilot will "fly" the simulation as if it's the real thing; you will have recreated the autopilot on a different piece of hardware. Similarly, if we can get a sufficient model of the brain to be able to simulate even most of its function, it will "think" in a similar way we do - even if it's in a machine and not in our head.

We've already done this for the brains of lesser animals. I recall a very accurate simulation of a flatworm's brain based on microscopic sectioning, and one pretty decent simulation of a fruit fly brain. The step from there to a human brain is a huge one, but it's just very hard - not impossible.

[nathaniel 0]

Quote

I think the biggest problem is not computer speed or I/O ports but how the hell you can even represent thoughts or feelings in a digital manner.

It's this line of inquiry that really appeals to me about the book Solaris by Stanislaw Lem. Before Hollywood ruined it with George Clooney.

Have we established yet that one can represent thoughts with thoughts?

Segmentation Violation

nathaniel

My advice is to do what your parents did; get a job, sir. The bums will always lose. Do you hear me, Lebowski?

[likearock 2]

Quote

We've already done this for the brains of lesser animals. I recall a very accurate simulation of a flatworm's brain based on microscopic sectioning, and one pretty decent simulation of a fruit fly brain. The step from there to a human brain is a huge one, but it's just very hard - not impossible.

I'd never say it wasn't possible. I just think that it's erroneous to point to the advances in computer technology and say we're on our way to AI. We would need a huge technological breakthrough before we could even start talking about representing consciousness in such a way that it can replicated and transmitted precisely, the way we can with images and sounds.

Wayne

[dorbie 0]

Quote

based on microscopic sectioning,

The step from there to a human brain is a huge one, but it's just very hard - not impossible.

Dunno about you but that sounds more than merely hard to me. Do we have a volunteer?

It reminds me of a comedy sketch or cartoon I saw once, on teleportation. You get into the teleportation booth get scanned and transmitted to the new location and recreated, voilla you're there. The jist of the joke was that scan was not destructive, the guy in the original booth was still standing there.... until the booth corrected the anomaly by destroying the original copy. In those last few seconds after the scan the passenger finally understands how a teleporter really works.

Booth: "Teleportation complete"

Passenger: "WTF? No wait!!!!! Ugh..."


Like I said read "The Emperors New Mind" by Roger Penrose. The strong AI & brain simulation proponents may not have all the answers.

[EricTheRed 0]

IMO by the time we create a true AI, our future as a species becomes rapidly irrelevant. A computer (or cyborg) mind would be able to engineer greater minds and the one thing that sets us apart from the rats and locusts will no longer be very useful.

I don't think any terminator type scenerio is probable but I certainly don't think that humans will be running the show for long after we create AI.

I also believe that it will happen. Don't have any guess as to when but I'd say that 50 years is totally possible.

There are people working very hard to make it happen.

http://www.singinst.org/

Edit to add:

http://research.microsoft.com/os/singularity/

illegible usually

[AMax 0]

Quote

Our understanding of the brain is like that - we can predict that if you damage the visual cortex in a certain place you will lose the ability to see certain kinds of patterns, but we're not always sure how each piece contributes to the whole.

Correct, even though we made a lot of progress understanding the role of each region, there is still a lot to be learned. Again connectivity between different modules is a biggest problem. Take stem sells. There are some fantastic examples of stem-cell induced regeneration of neurons damaged by spinal cord injuries, however there is a still huge problem – we cannot reproduce the exact pattern of innervation.

Quote

The problem isn't that we don't understand neurons, or even how they work together, because we do.

We are not there yet, but I have to agree that we know a lot about basic principles of synaptic transmission.

Quote

The problem we are facing is that we don't know the program. It's like understanding how computers work at a very deep hardware level, and then being asked to explain how an autopilot can land a 747 in a storm.

Exactly!

Quote

Similarly, if we can get a sufficient model of the brain to be able to simulate even most of its function, it will "think" in a similar way we do - even if it's in a machine and not in our head.

We cannot ignore the details because every basic element of network (synapse) is not constant. Synaptic plasticity (modulation of the strength of individual synapses) plays a huge role in neuronal network.

Quote

The step from there to a human brain is a huge one, but it's just very hard - not impossible.

Agree ... I just do not see it happening in near future.

[dorbie 0]

Update on brain simulation:

http://www.newscientist.com/article.ns?id=dn7470