On Consciousness

[steenkh]

Rupert Sheldrake has shown that telepathy is possible between people, especially between people who know each other well.

Has he now? There is a million bucks waiting for the first couple of people who know each other well who will demonstrate this ...

 

[Anders Lindman]

Has he now? There is a million bucks waiting for the first couple of people who know each other well who will demonstrate this ...

Sheldrake said that he will not accept the terms for the million dollar contest, because of things like copyright issues or something like that (I heard him talk about that but don't remember in which video).

 

[Mr. Scott]

the view expressed by the vast majority here is rationally necessary in the prevailing world view, but it is not supported by the evidence

As I've said before, there are a number of things people do easily that computer programs have great trouble doing...

...until they are modeled as neural networks.

Since the brain is made up of neural networks, I take this as evidence that the brain is a data processing machine that can be implemented in digital computers.

It doesn't bother me that it's taking longer than pundits or science fiction writers predicted. That's not evidence it can't be done.

What evidence have you that consciousness can never be implemented in a digital computer? "I can't imagine it" is argument from ignorance, not evidence.

 

[steenkh]

Sheldrake said that he will not accept the terms for the million dollar contest, because of things like copyright issues or something like that (I heard him talk about that but don't remember in which video).

Well, that only shows that he would rather not have his "scientific" results ruined by testing with real controls. He seems to know that it will not work in reality.

But anybody else who can do telepathy can grab the money, so when can we expect it to happen?

 

[Frank Merton]

As I've said before, there are a number of things people do easily that computer programs have great trouble doing...

...until they are modeled as neural networks.

Maybe; the history is not encouraging.

Since the brain is made up of neural networks, I take this as evidence that the brain is a data processing machine that can be implemented in digital computers.

A non-sequitur. Or maybe I should say circular argument. The idea of neural networks comes from trying to imitate brains.

It doesn't bother me that it's taking longer than pundits or science fiction writers predicted. That's not evidence it can't be done.

The same argument our friend below uses in favor of mental telepathy.

What evidence have you that consciousness can never be implemented in a digital computer? "I can't imagine it" is argument from ignorance, not evidence.

That is not fair. I made it perfectly clear that I just simply don't know: that my whole position is agnostic: that I just think the assurance on the other side is unjustified by any evidence.

Also, I resent your bringing up the argument from ignorance. I did not mention it.

 

[Anders Lindman]

Well, that only shows that he would rather not have his "scientific" results ruined by testing with real controls. He seems to know that it will not work in reality.

But anybody else who can do telepathy can grab the money, so when can we expect it to happen?

The kind of telepathy Sheldrake has shown seems real but very weak and unreliable (although independent of distance). I'm more interested in if telepathy could be used for forming a global mind where each person's brain and body develops a higher abstraction level that is a combined global nonlocal wave pattern correlation.

 

[steenkh]

The kind of telepathy Sheldrake has shown seems real but very weak and unreliable (although independent of distance).

No, it does not seem real at all. Weak and unreliable data are not evidence of something that is real. As a general rule, you can expect that Sheldrake will never come up with scientific results that can be trusted.

I'm more interested in if telepathy could be used for forming a global mind where each person's brain and body develops a higher abstraction level that is a combined global nonlocal wave pattern correlation.

Prove telepathy first, and then you can speculate on what it can be used for.

there is no evidence what so ever that the consciousnesses of our individual brains are part of a bigger consciousness.

 

[Mr. Scott]

Maybe; the history is not encouraging.

What, specifically, is the most discouraging example of this history I can read up on?

There's actually pretty good history of progress with AI. It's still in its infancy. Critics sound to me like people who would have said that, since we could only jump a few feet, we could never reach the moon. The human brain is the most complicated thing we know of. Understanding it and duplicating its function will take time. I've seen no evidence that it's impossible.

A non-sequitur. Or maybe I should say circular argument. The idea of neural networks comes from trying to imitate brains.

Maybe I skipped a step or two that seemed obvious to me. We see neural networks when we look in the brain. We found out how neurons behave. We see they are in networks. and see how the networks function. We simulate these networks in computers, and the result is we can make computers perform tasks that they couldn't perform before -- tasks that people previously beat computers at, thereby contradicting the claim that computers could never do the tasks brains could do because there's something special going on in the brain that's quite unlike data processing. Help me see the non sequitur or circular logic that you see.

That is not fair. I made it perfectly clear that I just simply don't know: that my whole position is agnostic: that I just think the assurance on the other side is unjustified by any evidence.

Also, I resent your bringing up the argument from ignorance. I did not mention it.

You are suggesting the brain is more than collections of neural networks. What evidence can you cite for that?

I have a very small kernel of agnosticism. That's why I started this tread. However, I argue what I think is the likely answer aggressively in the hope that it will shake out the truth. Sorry to offend you.

 

[dlorde]

In sum, the progress you talk about is quantitative, the difference I talk about seems to be qualitative.

I think both are necessary, and both are on their way. We're only just beginning to be able to create manageable systems with a number of computing elements comparable to even small mammalian brains (IBM recently ran a human brain size computational network, but it required all their available supercomputing facilities).

In the qualitative direction, we're now producing neuromorphic elements and connecting them in mapped architectures that follow how biological brains are connected. This is a qualitatively different approach to previous efforts.

I think there's a very long way to go, but optimistically, I expect to see an exact insect brain (probably a bee) emulation within 2 years, a rodent brain emulation within 5-7 years, and beyond that my crystal ball is cloudy...

 

[Wowbagger]

The idea is that consciousness is a single integrated entity that is aware of a huge amount of well-integrated information. That's what the Phi function is about. Phi is a value that gets higher the more such information is available to a single integrated system. The value of Phi can be low for a huge amount of information that is randomly connected. The information needs to be integrated is special ways, in addition to be plenty, for the value of Phi to be high.

I did not see the whole video, yet. But, based on my estimate as to what he is getting at, so far, I would respond this way:
At best, what he could be measuring is how reliable the information we are conscious of is, against the actual outside world around it. I do not think this approach can measure consciousness, itself.

Most of the filtering and such you are talking about does not go into conscious awareness.

Our understanding is no further along, nor our capabilities, than they were in my day.

I don't think this assessment is fair. We have a LOT more understanding about how various classes of AI work, such as neural networks, evolutionary algorithms, etc; and many more ways to work with them, and combine them, etc. We can devise ways to get AI working in ways similar to our own brains, at least in an abstract manner, and we improve upon those over time.

Is this enough for Strong AI? Probably not, yet.

However, there is no indication such a thing would be impossible for computing machines to have, yet, either. Until such a principle is discovered, I think the better attitude is to just keep trying! The more we learn about how consciousness works in the brain (and there is some science starting in that direction), the better chance we have to work out what to put into the computers to achieve the same effect.

I don't think Watson qualifies as anything close to a conscious machine. But, the claim that we are "not further along" than in your day, does not seem to be accurate.

 

[Wowbagger]

You are suggesting the brain is more than collections of neural networks. What evidence can you cite for that?

There are also hormonal-based "messaging" systems to consider, that have a significant impact on how the brain works and thinks.


Also, the way the networks developed could be critical. It's not enough to slap together a bunch of virtual neurons, and expect it to work like a human brain after many rounds of training.

To achieve human-like thinking, one probably also has to learn how human neurons evolved, and how they develop embriologically. Certain types of structures might emerge from there, that would be critical to human-like thinking, that would also be completely non-intuitive for humans to even guess at.

 

[dlorde]

Rupert Sheldrake has shown that telepathy is possible between people, especially between people who know each other well. Is that form of telepathy a nonlocal phenomenon? Yes. Is information being transmitted telepathically between people's minds? No. The nonlocal phenomenon of telepathy is a correlation function.

Not sure I follow - what is telepathy if not the transfer of information between two minds/brains by paranormal/extrasensory means?

If not information transfer, what did Sheldrake show (assuming he showed anything) ? linky? reference?

The human brain has trillions and trillions of waves of different frequencies. And when two people know each other well, their neural patterns are correlated so that their wave patterns share many frequencies.

I wouldn't be surprised if any two randomly selected people had some brain wave frequencies in common, but I'd be very interested to see evidence of greater correlation between people who know each other well than between strangers. Who is it that can measure these 'trillions and trillions of waves' and find correlations between them in individuals? Have you a reference or link?

 

[deaman]

In this thread we watch as consciouness argues with, and about itself.

 

[Anders Lindman]

Ray Kurzweil is skeptic about how useful artificial neural networks are. What I have seen is that in those networks the connections only go in one direction!

Is that really how the brain works? Aren't neurons connected to each other in both directions? Are synapses or what they are called only one-directional?

 

[Anders Lindman]

Not sure I follow - what is telepathy if not the transfer of information between two minds/brains by paranormal/extrasensory means?

As a crude (inaccurate) illustration, think of an incompressible rod one mile long. If two persons hold each end of the rod and one of them shakes the rod, the other person will sense the shakes instantly on the other end, without any time delay at all. In physical reality no such rod exists, but the sine waves that stretch the entire universe (see previous posts) work that way in actuality; when the amplitude of such sine wave is altered it is changed instantly and nonlocally along the entire length (or volume when considering 2D waves spanning all of space).

 

[steenkh]

What I have seen is that in those networks the connections only go in one direction!

Some simple neural networks only work in one direction, but it is not a problem to connect all of the neurons with everybody else, even though in a real brain, they are not all in communication with all of the rest.

 

[dlorde]

Is that really how the brain works? Aren't neurons connected to each other in both directions? Are synapses or what they are called only one-directional?

Pretty much. Neurons typically have a single 'output' axon and multiple branching 'input' dendrites where other neuron's axons connect.

At the majority of synapses, signals are sent from the axon of one neuron to a dendrite of another. There are, however, many exceptions to these rules: neurons that lack dendrites, neurons that have no axon, synapses that connect an axon to another axon or a dendrite to another dendrite, etc. (wikipedia)

 

[dlorde]

As a crude (inaccurate) illustration, think of an incompressible rod one mile long. If two persons hold each end of the rod and one of them shakes the rod, the other person will sense the shakes instantly on the other end, without any time delay at all. In physical reality no such rod exists, but the sine waves that stretch the entire universe (see previous posts) work that way in actuality; when the amplitude of such sine wave is altered it is changed instantly and nonlocally along the entire length (or volume when considering 2D waves spanning all of space).

I fail to see what that has to do with whether telepathy involves information transfer or not. Either it does or it doesn't. Which is it?

 

[Anders Lindman]

I fail to see what that has to do with whether telepathy involves information transfer or not. Either it does or it doesn't. Which is it?

Transfer implies something moved from one location to another location. That is different than the wave correlation in telepathy, where the wave changes amplitude along all positions in the universe at the same time.

 

[Anders Lindman]

For those unfamiliar with the Fourier transform here is a graph showing individual sine waves at the top and a combined wave of the sum of all the sine waves at the bottom: http://ars.els-cdn.com/content/image/1-s2.0-S0074614210970010-gr12.gif

From: http://www.sciencedirect.com/science/article/pii/S0074614210970010

Notice that the individual sine waves stretch the entire dimension and are "nonlocal" while the sum has local differences.