Materialists......

[nescafe]

Tell me, how do you disprove to your satisfaction solipsism? It is irrelevant as a useful position, we agree.

It is logically complete and consistent, and humans are not. Since solipsism as a philosophy is only applicable to one human (by definition), where does the completeness and consistency go?

The fact that it has no explanatory power beyond "well, I must have wanted it to happen" doesn't help either.

 

[UndercoverElephant]

Hold on there, cowboy. I do not agree with you at all that it is impossible, perhaps you missed post 528. Impossible with out current understanding of how neural networks work, sure. Impossible for all time -- I rather doubt it.

The problem is not that we do not understand how they work. We do understand how they work - that is the problem. It is exactly because we DO know that neural nets store data holistically/holographically that means we know we cannot extract partial information from them.

 

[nescafe]

The problem is not that we do not understand how they work. We do understand how they work - that is the problem. It is exactly because we DO know that neural nets store data holistically/holographically that means we know we cannot extract partial information from them.

Hmmm... so where is this proof that the current understanding we have of how neural networks work is the best understanding we can have?

 

[chulbert]

The information encoded within a neural net is distributed over the entire network. It is impossible to decode it at all. All you can do is observe the behaviour of the network and infer from this behaviour what sort of data it has encoded. But it is quite impossible to say exactly how or where that information is within the network because it is represented holisitically. None of it means anything in isolation from its relationship with the rest of the network, so you cannot retrieve anything at all from it unless you "ask" the entire network at the same time!

You are roughly correct about the inability to pinpoint information within a neural network but not for the reasons you think. The metaphor of a neural network that "encodes" or "stores" information is not entirely accurate because they are not a storage mechanism, they are a computational mechanism.

The inability to isolate "Madrid is the captial of Spain" in your brain is no different from the inability to isolate the point (4,16) in the fuction [latex]$$ f(x) = \frac{7}{2}x+2 $$[/latex]. It's not really "in there" so there's nothing to "decode".

Okay, okay, I just wanted to use the LaTeX tag.

 

[Arkan_Wolfshade]

The problem is not that we do not understand how they work. We do understand how they work - that is the problem. It is exactly because we DO know that neural nets store data holistically/holographically that means we know we cannot extract partial information from them.

If a neural net is storing the data holographically, then any point in the holograph can be retrieved without use of the neural net.

 

[Tricky]

You are correct.


See answer above. You should have stopped while you were ahead.

LOL. I still am.

I'd say your computer has as much compatabalistic free-will concerning its' "thoughts" as does any other machine. You apparently have faith you are a machine, so that includes you.

Is the presence of "compatablistic free-will the single criterion by which you judge the ability to think and presence of consciousness? If so, why don't you prove that you have it.

Do you consider your meaningless "thoughts" an epiphenomena? Or do you have some other explanation as to why it's irrelevant "noise" generated after-the-fact by the "meat machine" that is you.

Meaningless thoughts? Do you mean uncaused, or simply those that your memory cannot ascribe a reason for thinking. Funny thing about those "meaningless" thoughts. Sometimes you can backtrack on your chain of thoughts and figure out why you were thinking it. Or perhaps someone will say, something like "yeah, that song was just on the radio".

Or do you have some other explanation as to why it's irrelevant "noise" generated after-the-fact by the "meat machine" that is you.

Did I say anything about "irrelevant noise"? So pardon my interruption, but if this is the part where you tell me what my positions are, try to be clear so I'll know what I'm thinking. I just want to stay in the loop.

 

[UndercoverElephant]

If a neural net is storing the data holographically, then any point in the holograph can be retrieved without use of the neural net.

How? Maybe the analogy with a holograph isn't the best one....it wasn't actually mine. The point is that the information encoded in the neural net is encoded over the whole net. Small parts of the net simply do not mean anything when seperated from the rest of the net.

 

[UndercoverElephant]

It's not really "in there" so there's nothing to "decode".

Exactly.

 

[Belz...]

The data stored in your brain is stored in the same way that the neural net stores the data in needs to distinguish between patterns. The data in your brain is not stored in a symbolic "language of thought" of the sort proposed by Fodor several decades ago. "LOT" theories are also dead in the water.

Why can't I remember a specific person's name, sometimes, and sometimes I can ? It's really just a little bit of data, and yet it can be isolated.

 

[Arkan_Wolfshade]

How? Maybe the analogy with a holograph isn't the best one....it wasn't actually mine. The point is that the information encoded in the neural net is encoded over the whole net. Small parts of the net simply do not mean anything when seperated from the rest of the net.

With holographic data storage, the data is stored at a point where vectors intersect in a given n-dimensional matrix. The point can be accessed without knowing what conditions would cause the neural net utilizing said holograph to access the same point.

ETA: Sounds like you mean Storage Area Network, or Network Area Storage rather than holograph.

 

[UndercoverElephant]

Hmmm... so where is this proof that the current understanding we have of how neural networks work is the best understanding we can have?

Arghhhh!!! Understanding it better won't solve the problem. See chulberts post.

 

[UndercoverElephant]

Why can't I remember a specific person's name, sometimes, and sometimes I can ? It's really just a little bit of data, and yet it can be isolated.

You are talking about internal access. We all know we can do this. The debate is about external access. This we cannot do, for the reasons explained better by chulbert than by myself.

 

[Belz...]

YES!!!!!!!

The information encoded within a neural net is distributed over the entire network. It is impossible to decode it at all. All you can do is observe the behaviour of the network and infer from this behaviour what sort of data it has encoded. But it is quite impossible to say exactly how or where that information is within the network because it is represented holisitically.

So the designation of person "JustGeoff" is saved ALL OVER my brain ? Somehow, I'm sure there's a great number of ways my brain could get damaged and me NOT forgetting that, and vice-versa.

 

[Belz...]

We are discussing philosophy, and my comments pertain to this thread. What are you doing here, if you don't understand that?

My apologies, then.

 

[Belz...]

Perhaps, for some.

Tell me, how do you disprove to your satisfaction solipsism? It is irrelevant as a useful position, we agree.

Solipsism is unfalsifiable, so I can't. And yes, we agree.

The way I see it, materialism is the only useful starting assumption.

 

[hammegk]

LOL. I still am.

LOL indeed. Evidence?

Is the presence of "compatablistic free-will the single criterion by which you judge the ability to think and presence of consciousness?

No it isn't, and that you would think so demonstrates, again, how far out of your depth you are in this thread.

If so, why don't you prove that you have it.

I choose to believe, and act, as if I have actual "free-will". I have no interest in "proving" anything about this subject to you. Go read a book; google some; something.

Meaningless thoughts? Do you mean uncaused, or simply those that your memory cannot ascribe a reason for thinking. Funny thing about those "meaningless" thoughts. Sometimes you can backtrack on your chain of thoughts and figure out why you were thinking it. Or perhaps someone will say, something like "yeah, that song was just on the radio".

More meaningless babble.

Did I say anything about "irrelevant noise"? So pardon my interruption, but if this is the part where you tell me what my positions are, try to be clear so I'll know what I'm thinking. I just want to stay in the loop.

Tricky, you couldn't buy a clue. Why not go crap elsewhere on something else you don't understand, and have no interest in learning about.

 

[Belz...]

So you can't just read the disk sector then?

That's a surprising comment. If I read the disc sector directly, I'm not actually reading the DATABASE, am I ? And unless I know precisely what I'm looking for, which in a brain is unlikely, I'm just reading bits.

If the network has been trained to recognise whole faces, then you're stuffed.

I see. Well, I suppose this kind of architecture would be useful for some tasks but not much for another.

And I'm bored of responding to your posts, so from now on I'll just respond to anything you post that is actually worth responding to.

Oh, I'm sorry if I don't meet your standards of quality.

 

[Belz...]

Tricky, you couldn't buy a clue. Why not go crap elsewhere on something else you don't understand, and have no interest in learning about.

See, hammy. This is exactly why I don't like philosophy anymore. Arguing like this is pointless, because you guys simply STATE that you are correct without providing any form of reasoning to support it.

It's a good thing science doesn't work that way.

 

[hammegk]

It is logically complete and consistent, and humans are not.

Possibly relevant; possibly not.

Since solipsism as a philosophy is only applicable to one human (by definition), where does the completeness and consistency go?

Why do you consider it only applicable to "one human"?

The fact that it has no explanatory power beyond "well, I must have wanted it to happen" doesn't help either.

Nor I am sure who or what that "I" is, but we are in basic agreement otherwise.

Arguing like this is pointless, because you guys simply STATE that you are correct without providing any form of reasoning to support it.

You don't know enough to argue. Kindergarten is in a different thread.

 

[.13.]

The answer to that is not available until the relationship between subjective and objective has been deconstructed. But it is not deconstructable until that relationship has been properly recognised and understood. And it is not properly recognisable or understandable so long as people keep on trying to defend materialism instead of trying to understand the true relationship between subjective and objective.

I see. Since you can't say what other necessary conditions there are I'll have to conclude that brains are in fact sufficient for minds.

So why do you think then that brains are necessary but not sufficient conditions for minds to exist?