Merged Cognitive Theory, ongoing progress

[barehl]

Let me see if I can clarify this.

If I had a grant to do experimental research then credit wouldn't be much of an issue. Whatever conclusions I came up with would have to be supported by the experimental evidence. Someone could try to publish the conclusions but unless they had copied the data, procedures and results they would have no evidence to back it up. And, if they did copy these things you would still have your lab notebook as evidence. It would be very difficult for someone to steal your research and try to take credit for it. I assume this is what Reality Check was talking about when he mentioned 'baseless fears'. Unfortunately, this isn't what I'm doing.

I'm trying to create a high level model of the brain. I'm not doing basic research and conducting experiments. There are already vast quantities of information available that no one has been able to interpret. I look through the same evidence and try to find patterns. The patterns form hypotheses and these are then tested against other evidence. Hypotheses that fit with the evidence are included in the model. Then I look for more evidence that will support or disprove the model. This is what Charles Darwin did to create evolutionary theory. It wasn't based on laboratory experiments; it was based on interpreting existing evidence. The problem though is that if your ideas are based on existing evidence then only the ideas themselves have value. It was only the threat of being scooped by Wallace and losing credit that forced Darwin to publish. The great majority of people working on cognitive theory gave up on the idea of a high level model years ago. The current effort is based on either extended AI (which hasn't worked) or neural networks (which hasn't worked) or neural modeling which failed spectacularly in the Blue Brain Project. So, while I'm far from the first to work on this I don't know of anyone else who is still doing it.

So rather than being childish and daring me to provide more detail, someone would need to explain why credit concerns would not be valid. No one has even attempted to do that. If you actually believe that it isn't a concern then it should be easy for you to explain why. Just saying that the concerns are baseless is not much of an explanation.

Secondly, I have great concerns about publishing a theory like this with an obviously, incompetent fool like Trump in the White House, and Republican majority control of the House and Senate. So, I'm not anticipating publishing until these things change. No one has attempted to explain why this also shouldn't be a concern.

 

[fagin]

So you've got nothing?

 

[Myriad]

Secondly, I have great concerns about publishing a theory like this with an obviously, incompetent fool like Trump in the White House, and Republican majority control of the House and Senate. So, I'm not anticipating publishing until these things change. No one has attempted to explain why this also shouldn't be a concern.

There can be no attempt to explain why this shouldn't be a concern, because you haven't explained the nature of your concern.

Do you fear government reprisals for publishing a theory of cognition that the President or Congressional majority might (piling speculation upon speculation) disagree with?

Are you hoping that publication will result in government funding or other support for your work, which would less likely be forthcoming from the current administration than from some future one?

Are you concerned that the Republican Party might misuse your findings, to the detriment of the public? (If so, couldn't they and other questionable political regimes around the world any time subsequent to publication, misuse it anyhow even if they're not in power at the time of publication?)

Is it your intent to punish the world for permitting the U.S. to elect an executive and representative body that you disapprove of, by withholding your valuable discoveries?

I speculate the above four possibilities because you have not specified a reason. My purpose is not to attempt to win a guessing game but to demonstrate where the available information falls short of sufficient clarity for a coherent rebuttal. None of those reasons seem very plausible to me, but I haven't been able to concoct any others that are any more so.

 

[barehl]

Do you fear government reprisals for publishing a theory of cognition that the President or Congressional majority might (piling speculation upon speculation) disagree with?

Well, obviously Pubs will immediately disagree with it because it refutes life after death and is based on Evolutionary theory (clearly a tool of the Devil). I think condemnation from evangelicals and religious fundamentalists is a given. Death threats from Trump supporters wouldn't be out of the question.

But, let's say that the theory is correct. We could expect Russia, China, Japan, South Korea, India, and western Europe to jump on it. I couldn't expect any kind of coherent response from Trump so the US would be delayed, possibly by years. Nor would other Pubs (who routinely kiss the ring of the president of Liberty University) be eager to spend money on something anti-Christian.

I suppose some would argue that greed would win out but some elements of the theory don't favor big business.

 

[Pixel42]

I've mentioned a number of things about it and I could mention more. People here complain when I do.

You need to tell us enough for it to be possible to have a meaningful discussion of what you've told us. The only complaints I've seen are complaints that you haven't done that. If you don't feel you can do that that's fine, but you can't then be surprised that we aren't discussing things you haven't told us.

From what you are saying, you won't value my ideas until I am no longer here to discuss them. That does seem very odd to me.

No, I'm saying we can't value your ideas if we don't know what they are. And I don't understand why you could not return to discuss them after you've published them and we finally have something to discuss.

 

[Roboramma]

Hypothesis: human cognition and intelligence evolved over time from non-cognitive and lower intelligence organisms. To avoid lots of random, lucky jumps in brain structure, elements that are used at each level must be present in previous generations. And more than likely these elements will still be found in modern organisms that have lower cognition/intelligence.

I think to some extent that's quite reasonable and likely to be true.

I don't think it's a particularly original revelation, however. For instance, it bears considerable resemblance to the Triune Brain model:

https://en.wikipedia.org/wiki/Triune_brain

The triune brain is a model of the evolution of the vertebrate forebrain and behavior, proposed by the American physician and neuroscientist Paul D. MacLean. MacLean originally formulated his model in the 1960s and propounded it at length in his 1990 book The Triune Brain in Evolution.[1] The triune brain consists of the reptilian complex, the paleomammalian complex (limbic system), and the neomammalian complex (neocortex), viewed as structures sequentially added to the forebrain in the course of evolution. However, this hypothesis is no longer espoused by the majority of comparative neuroscientists in the post-2000 era.[2] The triune brain hypothesis became familiar to a broad popular audience through Carl Sagan's Pulitzer prize winning 1977 book The Dragons of Eden. The theory has been embraced by some psychiatrists and at least one leading affective neuroscience researcher.[3].

 

[Reality Check]

Let me see if I can clarify this.

Followed by no clarification.
Grants do not establish credit for work
What you imagine you are doing is nice but irrelevant unless you can give a coherent description of the status of this "high level model of the brain".
Describing or publishing science has nothing to do with who is president, etc. Donald Trump does not control every forum in the world. Donald Trump does not control every journal in the world.

 

[Reality Check]

Well, obviously ....

nothing is ever published on evolution because "condemnation from evangelicals and religious fundamentalists is a given. Death threats from Trump supporters wouldn't be out of the question." !
You need to have a look at my user name and check the reality of the many evolutionary theory papers, articles and forum posts that have been written, even in the last year.

 

[barehl]

I don't think it's a particularly original revelation, however.

Why would it be? It seems to me to be a fairly obvious idea from evolutionary theory. Is that what you think I was referring to as my theory?

For instance, it bears considerable resemblance to the Triune Brain model:https://en.wikipedia.org/wiki/Triune_brain

I'm not seeing any connection. For example I am not trying to map sections of the brain. Secondly, I don't agree with his generalizations. If you think it has "considerable" resemblance you'd have to explain to me where.

 

[barehl]

You need to tell us enough for it to be possible to have a meaningful discussion of what you've told us.

I listed what I currently think about general AI and ran through the types that I had investigated. There was not a single response to that. So, what is it you want me to tell you?

are complaints that you haven't done that.

That I haven't done what?

And I don't understand why you could not return to discuss them after you've published them and we finally have something to discuss.

That seems highly unlikely.

 

[barehl]

To be honest, I'm having trouble figuring out who or what you think you are responding to. You seem to believe that you are replying to me, but I'm not seeing much connection between what you wrote and what I've posted.

Grants do not establish credit for work

Why would they? I never said they did.

What you imagine you are doing is nice but irrelevant

So, why are you involved with this thread?

unless you can give a coherent description of the status of this "high level model of the brain".

What do you mean by "status"?

Describing or publishing science has nothing to do with who is president, etc.

Yeah, that's obvious.

Donald Trump does not control every forum in the world.

Again, obvious.

Donald Trump does not control every journal in the world.

And, again, obvious.

None of your statements add anything; they were a complete waste of time. You did have one question which I'll try to answer if you can explain what you are asking for.

 

[Roboramma]

Why would it be? It seems to me to be a fairly obvious idea from evolutionary theory. Is that what you think I was referring to as my theory?

I understand that it's not the whole of your theory. I also think it's, as I said, quite a reasonable conclusion of evolutionary theory, though not one that I think would hold true always, I think it's a good general principle.

However, you were asked about evidence consistent with your theory and you presented this, it seemed that the implication was that there was something interesting here: that this was evidence specifically connected to your ideas and not simply to ideas that are already well understood in evolutionary theory. Did I misunderstand?

I'm not seeing any connection. For example I am not trying to map sections of the brain. Secondly, I don't agree with his generalizations. If you think it has "considerable" resemblance you'd have to explain to me where.

The connection is simply that he presented a theory based on the idea that modules of the brain evolved sequentially over evolutionary time and that evolution can be traced through our ancestry, which seemed to be the same idea that you were presented. I wasn't suggesting that it was the same in the specifics. I don't know the specifics of your idea, but I wouldn't expect you to have been presenting the same specific model as the Truine Brain, and I apologise if I gave you that impression.

My point was the the general idea that you did present is also present in the Triune Brain model.

 

[Pixel42]

I listed what I currently think about general AI and ran through the types that I had investigated. There was not a single response to that.

Around here that usually means that what you've posted is not news and no-one particularly disagrees with it.

What response were you expecting?

So, what is it you want me to tell you?

Anything that might actually prompt a discussion. Around here that usually means something original and/or controversial.

That I haven't done what?

Told us enough for it to be possible to have a meaningful discussion of what you've told us.

That seems highly unlikely.

Why?

 

[barehl]

However, you were asked about evidence consistent with your theory and you presented this, it seemed that the implication was that there was something interesting here: that this was evidence specifically connected to your ideas and not simply to ideas that are already well understood in evolutionary theory. Did I misunderstand?

Is intelligence an advantageous trait?
If it is then why isn't everything intelligent?
Why were there no dinosaur civilizations?
Where do abstractions come from?
What gives abstractions value?
How do humans understand things?
Why can't computers understand things?
Could humans be super smart someday?
Could we encounter aliens vastly smarter than us?
Could a computer suddenly become conscious?
Could a network become conscious?
How much processing power would be required to duplicate the human mind?
How much memory?
Why do people have emotions?
Can you build a box with a human personality like in the movie, "Her"?
How does the brain overcome the frame problem?
How does the brain solve the binding problem?
Can we build smart and loyal robots?
Could singularity happen?
Could you upload your mind to a computer?
What is awareness?
Is it the same as consciousness?
How does problem solving work in the brain?
Is Hofstadter's feedback loop model correct?
Is Integrated Information Theory correct?
Is Dennett's Multiple Drafts model correct?
What about the Chinese room?
What about Mary's room?

These are the types of questions that concern me.

The connection is simply that he presented a theory based on the idea that modules of the brain evolved sequentially over evolutionary time and that evolution can be traced through our ancestry, which seemed to be the same idea that you were presented.

That's basically what Darwin said. That isn't a significant point.

 

[barehl]

Around here that usually means that what you've posted is not news and no-one particularly disagrees with it.

Well, I know for a fact that some here disagree with it.

What response were you expecting?

A response.

Anything that might actually prompt a discussion. Around here that usually means something original and/or controversial.

I'm the only person who has ever claimed to be able to refute Idealism. I guess that isn't original enough. I know what mistakes Kurzweil made in his estimate of hardware to duplicate the brain. I don't know of anyone else who does. I'm the only one working on a high level model. Not original. Interesting claim.

Told us enough for it to be possible to have a meaningful discussion of what you've told us.

I still don't know what this means.

Why?

Because I would almost certainly be discussing it elsewhere.

 

[Pixel42]

Well, I know for a fact that some here disagree with it.

Then they either didn't disagree enough to think it worth arguing about, hadn't seen the thread, or didn't think an argument with you would be fruitful.

A response.

What response? What did you expect people to say?

I'm the only person who has ever claimed to be able to refute Idealism. I guess that isn't original enough. I know what mistakes Kurzweil made in his estimate of hardware to duplicate the brain. I don't know of anyone else who does. I'm the only one working on a high level model. Not original. Interesting claim.

Yes, that would merit considerable discussion. But you not only didn't post any of it, you declared you had no intention of posting any of it. So what was there to discuss?

I still don't know what this means.

Sorry, I can't put it any more clearly.

Because I would almost certainly be discussing it elsewhere.

And it would then be beneath you to visit this forum. I knew that was the explanation, I just wanted you to come right out and say it.

 

[Beelzebuddy]

Look, you are talking to academics about this at least, right? At the very least I'd strongly recommend making friends with cognitive scientists at your local uni. Send them an email, buy them some coffee. You'll find them more receptive than you think.

No, I'm not. There aren't any. I've probably sent out two dozen [emails]. I did get a reply from one who said that he would be afraid of a machine with human intelligence. [Finding them receptive] hasn't happened yet.

Yet:

Please stop being dishonest.

So rather than being childish and daring me to provide more detail, someone would need to explain why credit concerns would not be valid.

None of your statements add anything; they were a complete waste of time.

Gee I wonder why.

I listed what I currently think about general AI and ran through the types that I had investigated. There was not a single response to that.

"If you don't have anything nice to say..."

I, at least, have been waiting for you to get through your manifesto to what you actually think is new. So far you haven't said anything that couldn't be found in a textbook, more often than not as an example of how alluring wrong ideas can be. If this is how the rest of it goes as well, you don't need to worry about credit because the people who deserve the credit are already dead.

 

[W.D.Clinger]

And I don't understand why you could not return to discuss them after you've published them and we finally have something to discuss.

That seems highly unlikely.

It does indeed seem highly unlikely that you will publish your results and give us something to discuss.

These are the types of questions that concern me.

Okay, I'll play your game.

Is intelligence an advantageous trait?

At times it may be.

Stupidity may also be advantageous, at times. If it weren't, why do we see so much of it?

If it is then why isn't everything intelligent?

When bacteria rule the world, it makes more sense to ask why isn't everything stupid.

Why were there no dinosaur civilizations?

A lot of dictionaries define "civilization" as something like "The stage of human social development and organization which is considered most advanced." That, by definition, rules out dinosaur civilizations.

That same dictionary goes on to give this example: ‘the Victorians equated the railways with progress and civilization’. For all we know, some of the dinosaurs equated their nests and trails with progress and civilization. Who are we to argue?

Where do abstractions come from?

Depends on the abstraction. Your question may have been too abstract.

What gives abstractions value?

I've been told that the value of something comes from individuals' willingness to exchange or pay for it, but this is the wrong subforum for extended discussion of that.

How do humans understand things?

Poorly.

Why can't computers understand things?

My computer understands many sequences of bits a lot faster and more reliably than I do.

Could humans be super smart someday?

It might depend on the human.

Could we encounter aliens vastly smarter than us?

Sure. Whether we should encounter them is often discussed in the US Politics subforum.

Could a computer suddenly become conscious?
Could a network become conscious?

That might depend on what you mean by "conscious". In the past, I have asked you what you meant by that word, but I never understood your answer. As I recall, that discussion did not turn out well.

How much processing power would be required to duplicate the human mind?
How much memory?

Some people have tried to answer those questions, but I don't think we know enough about the human mind to answer those questions at this time.

Why do people have emotions?

Partly because their brains, including identifiable regions of their brains, have evolved to create emotions. We believe emotions are evolutionarily advantageous because some people who are abnormally unemotional suffer from it and may be classified as mentally ill.

Can you build a box with a human personality like in the movie, "Her"?

I have enough trouble just building a box. Besides, I haven't seen the movie. Do you recommend it? Why? (No spoilers, please.)

How does the brain overcome the frame problem?

Ah, the frame problem. I vaguely recall Marvin Minsky going on about that in an AI course I took long ago.

My personal opinion is that the frame problem is primarily relevant to expert systems and applied artificial intelligence. Brainful creatures generally avoid resort to logic until it becomes obvious that habit, training, prejudice, emotion and other preferred tools aren't getting the job done. Even then, most people have trouble using logic.

(I'm generalizing from my students, of course, but they've been selected in part for their ability to employ logic, and they've taken courses that are supposed to teach them logic.)

How does the brain solve the binding problem?

I believe most of the experimental research has involved vision, but I don't feel qualified to discuss that research in detail.

Can we build smart and loyal robots?

We've built cell phones, and those cell phones have created loyal owners. If some of those owners are smart, we can classify them as robots and declare victory here.

Could singularity happen?

To a mathematician, that's like asking "Could zero happen?"

If you're talking about what the popular press sometimes refers to as an AI singularity, that's a concept whose value lies primarily in the revenue it generates for the popular press.

Could you upload your mind to a computer?

I doubt it, but I've never tried.

What is awareness?

Huh? Oh. It might be related to attention and comprehension.

Is it the same as consciousness?

That might depend on what you mean by "consciousness", in which case the question will remain unanswerable until you explain what you mean by "consciousness".

How does problem solving work in the brain?

For the most part, brains solve problems similar to problems they have solved before, mostly by employing tactics that have worked before.

Most people who ask that question in the context of AI are asking about very artificial sorts of problems. Any good answer to that question will start by noting that those are the sorts of problems people seldom have to solve, often fail to solve, and normally have trouble with even when they do eventually arrive at a solution.

Is Hofstadter's feedback loop model correct?

It's been years since I last talked to him, so I am not familiar with the current details. In general, though, the correct answer to questions of the form "Is model X correct" is "Not entirely."

Is Integrated Information Theory correct?
Is Dennett's Multiple Drafts model correct?

See above.

What about the Chinese room?
What about Mary's room?

Mary's is not the room that interests me.

I am aware that some philosophers delight in finding clever ways to define problems in such a way that the problem, by definition, has no solution, and then use an obfuscated form of equivocation to get people to think they've made some kind of profound contribution to the theory of knowledge or intelligence.

 

[Beelzebuddy]

How does the brain solve the binding problem?

I believe most of the experimental research has involved vision, but I don't feel qualified to discuss that research in detail.

Hierarchical topographic maps. The binding problem was only a problem in the context of the hypothetical computer architecture doing the binding, which had to pick between global concepts devoid of spatial context, or localized information with no unifying architecture. Turns out nested topologies can translate between the two just fine.

If you're familiar with deep learning, it operates on a similar principle.

 

[W.D.Clinger]

Thanks.