Discussion Topic |
|
This thread has been locked |
WBraun
climber
|
|
|
Jun 23, 2017 - 05:55pm PT
|
The AI guys in 1964 were robots in consciousness so they couldn't understand anything beyond that and still can't to this day ......
|
|
Largo
Sport climber
The Big Wide Open Face
|
|
|
Topic Author's Reply - Jun 23, 2017 - 07:01pm PT
|
While I don't go with John Searl's causal/mechanistic view of consciousness - at least not entirely - he's gone some distance in teasing apart the differences between machine registration of an input, mechanical processing, and consciousness. If nothing else, Searl is always lucid. And like Uncle Dennett, he's a crusty old curmudgeon full of wit and kinetic energy.
http://static.trogu.com/documents/articles/palgrave/references/searle%20What%20Your%20Computer%20Can%E2%80%99t%20Know%20by%20John%20R.%20Searle%20%7C%20The%20New%20York%20Review%20of%20Books.pdf
And Healje, you know perfectly well that once he strays outside straight-up neuroscience, Markham is a total sham, the P.T. Barnum of "consciousness studies" who knows a sucker when he sees them and asks for a cool billion - and gets it - while talking about machines that will fall in love by 2020. He's selling snake oil in regards to consciousness.
|
|
yanqui
climber
Balcarce, Argentina
|
|
|
Jun 23, 2017 - 07:15pm PT
|
I'll see what I can learn but I may not be able to offer much. .
I'll look at this too. I don't think I have any problem with "emergence" (I haven't really thought about this much). I know a little about algebraic topology and also about graphs for neural networks. I studied something called "group representations" but my office mate (an Argentine named José) got interested in a certain kind of graph (called an arrangement graph) that has been studied a lot, apparently, for its relevance to neural networks. Anyways, José, who's a pretty smart guy (in the classical mathematical sense of the word) realized that a problem that some other guys were working on (called the spectral theory of the graph - a theory that completely determines the graph) could be solved using the ideas from group representations. I don't publish hardly ever, but I thought this was a really cool idea, so I had to learn about graphs and such. It turns out José was right and we gave a simple and elegant argument for the solution that was based on ideas that were 100 years old. One of the formulas we used comes from a text by Frobenius in German that was published in the 1890s. I love it when that happens, even if referees and editors and aren't really sure how to deal with it (at any rate, the result got published).
|
|
WBraun
climber
|
|
|
Jun 23, 2017 - 07:19pm PT
|
while talking about machines that will fall in love by 2020.
Hahaha lol
And idiots WILL buy this bullsh!t for sure .......
|
|
MH2
Boulder climber
Andy Cairns
|
|
|
Jun 23, 2017 - 07:22pm PT
|
That's what I like to hear, yangui
It was fun to see this paper reference Hebb (1949).
http://journal.frontiersin.org/article/10.3389/fncom.2017.00048/full
briefly:
Specializing basic concepts of algebraic topology, we have formulated precise definitions of cliques (simplices) and cavities (as counted by Betti numbers) associated to directed networks[of neurons].
I found no particularly interesting empirically testable consequences mentioned in the paper. There is the weak sense in which other researchers can use the simulation to look for consequences they might be interested in, but those would be model-specific.
There is a mention toward the end of how the cliques and cavities might provide for recognizing features in a stimulus and for making associations among different features. That would be much more interesting, and all that needs to be done is present the model with stimuli and see what happens.
A major problem, it seems to me, is that there is still a long distance between the kind of stimulus you could present to such a simulated network, and a stimulus a human might take interest in, like an image of a hammer.
The notion of stimulus is too broad. Context is needed.
If I come in the room and there's a dangerous intruder, then maybe I should see the hammer as a potential weapon. If I come in the room when I need to hang a picture, then I should see the hammer as a tool.
I congratulate you on the ability to recognize that you have entered a room. That is no small feat.
One of the many problems with so-called AI comes from our having no idea how hard apparently simple activities like walking or breathing are. Our measure of difficulty can’t be just our own subjective impression. I’m not aware of ANY good measure of the difficulty of such problems.
One of the photos I took on the trip to Smith Rock showed a colorful bird. I was curious about what kind it was. I used Google image search, which immediately recognized the image as a bird. I’m sure that current technology can do better, but more work is needed. On the other hand, my brother told me he used Picasa to go through his photos for pictures of one of his 5 children and that it did a pretty good job.
|
|
jgill
Boulder climber
The high prairie of southern Colorado
|
|
|
Jun 23, 2017 - 09:16pm PT
|
These mathematical models seem to reach points of intractable computations. Although, having little to no knowledge of algebraic topology - point set topology was as far as I wanted to go - I know nothing of computational problems to be encountered.
On the other hand, Tononi's Phi function in IIT has problems: "Rather than try to start from physical principles and arrive at consciousness, IIT "starts with consciousness" (accepts the existence of consciousness as certain) and reasons about the properties that a postulated physical substrate would have to have in order to account for it . . . While the algorithm[5] for assessing a system's Φ Max and conceptual structure is relatively straightforward, its high time complexity makes it computationally intractable for many systems of interest." (Wiki)
Ed, when you evaluate path integrals what are the computational problems, if any? Is good software available? I'm piddling with elementary functional integrals now and having no problems, but I'm curious about your applications.
|
|
healyje
Trad climber
Portland, Oregon
|
|
|
Jun 24, 2017 - 03:41am PT
|
Largo: And Healje, you know perfectly well that once he strays outside straight-up neuroscience, Markham is a total sham, the P.T. Barnum of "consciousness studies" who knows a sucker when he sees them and asks for a cool billion - and gets it - while talking about machines that will fall in love by 2020. He's selling snake oil in regards to consciousness.
Markham is a lot of things but he's not a sham. The link I just posted is to a just published paper from Markram's HBP project - focus on the work, not the hype that sold it. No one calls Kurzweil a sham and he's way the f*#k farther out there with half of what comes out of his mouth than Markram. Markram's biggest problem is he has none of the after-the-sale management capabilities of a Venter (not necessarily pleasant capabilities, but one which are ultimately effective even if at a human cost). And your tendency to latch onto relatively insignificant things like a dog on a bone and present them like they're the main attraction in an attempt to make your points is weak - all the moreso when you basically don't know what you're talking about.
When you talk about “magic,” I don’t think you have much of an idea what you’re talking about other than a connotation, which is pejorative. Anthropologists and depth psychologists have a much clearer notion of what magic is. They relate it to a more tribal sense of consciousness: being as an appendage of the tribe; pars pro toto, totum pro parte; where everything is connected to everything else; where all of nature is alive and numinous, where a symbol (think very old artifact) is more real than what we moderns think is real, etc. There autonomy and ego are barely beginning to emerge in Man.
That first sentence I'll respond to, the rest of the paragraph is on a journey of it's own without the slightest connection to my point. Let's stick with: "magia as an elemental force pervading many natural processes" which is what Largo is suggesting in terms of a fundamental and pervasive awareness. Again, which is more likely: consciousness emergent from brians or a convergence of brains and magia?
I’m not here to argue with the theory, but I would have folks give some consideration to other things that affect the changes in Man . . . like culture (which is complex) as well as learning. It is perhaps no coincidence that culture has developed across the earth in different times, and those developments are difficult to connect in space time. That is, culture (for example) seems to have emerged wherever man has been.
Learning is a builtin behavior of higher animals (squirrels are good learners as are octopi) and culture is simply the result of collective behavior of social animals. Our base behaviors are genetically based and played out, expressed within, and influenced by our environmental context. They are evolved behaviors. And behavior does play into evolution in a kind of feedback loop - that's where discovery, learning, tool use, etc. come into play and bear on our future. So can culture influence evolution? Sure, but that's clearly part of the evolutionary process for higher animals. Cat's are an example that comes to mind - they domesticated themselves. Also, with regard to your affinity to and concerns about 'values', it's worth keeping in mind that life began billions of years ago, the first brain structures in worms about half a billion years ago, and brains per se didn't evolve to the point of even 'having' values until quite recently - like a few picoseconds ago on evolutionary scales.
Last, I find myself wondering just what problem or challenge that consciousness solves in life. If random mutation allows for the selection of the characteristics that lead to (successful reproduction) in advanced life, then it would seem that consciousness would hardly be random. It would seem to me, at least, that it becomes imperative and necessary. But of course, I’m speculating.
Survival pure and simple. As far as consciousness being random, I'd say you're thinking about it the wrong way. Again, looking at extant species from bacteria to humans, there is a clear progression of capabilities from basic sense/response capabilities in bacteria on up through to the stalking/hunting behavior of both Jumping Spiders and Humans. It isn't like humans randomly and suddenly mutated to have a brain, but rather billions of years of steady evolution starting with the first sense/response capabilities and steadily improving through eons of species until central nervous systems and brains emerged and hundreds of millions of years later hominids eventually followed suit. Sense/response survived because it conveyed an advantage; it continued to evolve into brains, awareness, sentience and consciousness for the same reason. And as I keep saying, I believe ever more sophisticated predation has been the primary driver of that evolution across time and species.
Also, your portrayal of 'evolution' as an answer so broad it says nothing about consciousness is entirely wrong. The title of the thread is 'What Is Mind?" and the discussion has literally been around what it is and its origin as opposed to the universe of things we do with it or various values one might try to associate with it. Saying evolution is the reason you have a mind at all and mind is emergent from the meat is one of only a couple of fundamental stances within this discussion. And those discussion stances have pretty much come down to a) we have an evolved brain from which consciousness emerges or b) brains and a fundamental awareness somehow manage to converge to provide consciousness. I'd love to hear c), d) or e) ideas / theories, but for the moment a) and b) appear to be what we have.
|
|
Dingus McGee
Social climber
Where Safety trumps Leaving No Trace
|
|
|
Jun 24, 2017 - 04:08am PT
|
From the Wiki Phi function:
https://en.wikipedia.org/wiki/Integrated_information_theory
The properties required of a conscious physical substrate are called the "postulates," since the existence of the physical substrate is itself only postulated (remember, IIT maintains that the only thing one can be sure of is the existence of one's own consciousness).
Then a distinction would seem to exist between knowing the existence of one's own consciousness and being conscious of something? Does the IIT PHI math show this?
The PHI measure is interesting but I have not figured out or worked through the math to see what makes PHI increase. That is to say what is PHI differentiated?
As for the conscious required in bump skiing, all of the causal mental processes would be physical substrate and when the situation arises that I am conscious of doing bump skiing that consciousness is sufficient and necessary to have the physical substrate to bump ski. OKAY. When do we max PHI? There is a little talk on PHI max.
|
|
healyje
Trad climber
Portland, Oregon
|
|
|
Jun 24, 2017 - 04:49am PT
|
With regard to formalities, be they philosophical, math-based or some other form of science. It's all good, but as a starting point you pretty much have to lay out your assumptions (like that consciousness exists in IIT). For me the fundamental assumption is brains and consciousness exist and so the starting question is the one Largo answered: can consciousness exist without a brain? The answer is either yes or no and each takes you down a very different path in search of an answer to this thread's title question.
|
|
Dingus McGee
Social climber
Where Safety trumps Leaving No Trace
|
|
|
Jun 24, 2017 - 06:05am PT
|
MikeL
some behaviorist argue that our ethics and values are simply a statement of the way we did evolve to get along with each other. These values were not created by culture but are the summary worded expression of how we evolved to get along smoothly. The deviants [from some norm] act otherwise.
|
|
MikeL
Social climber
Southern Arizona
|
|
|
Jun 24, 2017 - 07:03am PT
|
healyje,
We’ll agree that we disagree. I’m probably too academic and picky about how I am willing to employ theories. My training compels me to be more specific about all sorts of things implied in research studies. “Learning,” for example, is impossibly broad. For example, some colleagues who worked for the Center for the Study of Reading at UIUC showed me that the theories / research around how children comprehend what they read is deep and complex (see http://csr.ed.uiuc.edu/publications.html);. I got my Ph.D. there, and the cognitive science group was pretty involved with those folks. Interesting work, and pretty specific stuff. No one ever argued for evolutionary theory that I remember.
Dingus,
Behaviorists ignore the mind; it’s a black box for them. They don’t think they need to know what’s going on in the mind or the brain. It’s been wondered why they ever got involved in studies of cognition and mind to begin with.
No one doubts that cooperation and competition work hand-in-hand, especially in business (so-called “coopetition).
If anything that shows up today *must be* the result of evolution, then there’s not much more to say or research. Evolution would apply to and explains everything about life, wouldn’t it? How could anyone dispute it if used in this fashion?
The theory of evolution is a very clever narrative. The conclusion is built into the premise.
|
|
yanqui
climber
Balcarce, Argentina
|
|
|
Jun 24, 2017 - 07:09am PT
|
Our measure of difficulty can’t be just our own subjective impression.
All I can say about this is that even in the formally defined realm of computational problems (computational complexity) this is often the only measure there is. You could win a million dollars if you could solve a small part of this problem (in the context of computational complexity) by doing this:
http://www.claymath.org/millennium-problems/p-vs-np-problem
|
|
WBraun
climber
|
|
|
Jun 24, 2017 - 07:14am PT
|
P vs NP Problem
This problem is: Unsolved
But!!!! The scientists Believe and have FAITH that in the future, "THEY" WILL solve it ......
|
|
MH2
Boulder climber
Andy Cairns
|
|
|
Jun 24, 2017 - 07:17am PT
|
Is there anyone smart out there?
Is it enough to be conscious, or aware of the existence of consciousness?
Is Google conscious?
This bugs me, but I may not smart enough to do a good image search:
Did software REALLY beat Gary Kasparov at chess, or was that a sham?
|
|
MH2
Boulder climber
Andy Cairns
|
|
|
Jun 24, 2017 - 07:35am PT
|
even in the realm of formally defined computational problems (computational complexity) this [a subjective impression] is often the only measure there is.
If you need to lift an object you can't lift by yourself, you might solve the problem by getting enough people together to help lift it. When the summer camp put its floating dock in the water it took about 50 people around the perimeter.
Problems that require thinking rather than physical effort are different. Much smarter people than me have put mental effort into the search for general problem-solving methods.
The guy who introduced me to rock-climbing was a physics student at the time but he became a computer scientist. I once asked him if the then new [mid-80s] developments in parallel processing would be important to him. His reply was that he did not work at the level of architecture.
I used to think math was abstruse, until one day in the University of British Columbia library I found myself in the computer science section and although the articles seemed to be written in English I could not understand a single sentence.
|
|
yanqui
climber
Balcarce, Argentina
|
|
|
Jun 24, 2017 - 07:58am PT
|
can consciousness exist without a brain?
In this case my belief is no and I would be highly unlikely to accept any theory of consciousness that entailed this. Come to think of it, it seems to me that the language of probability could be the best language for rational beliefs. Has anyone (say a philosopher or scientist) ever talked about that before? But I digress.
Edit to make point clearer: Of course we can think abstractly about consciousness as something apart from the brain. In a way Turing did this, to some extent when he formalized one kind of activity that mathematicians do consciously (and subconsciously) when we use formulas to calculate.
However, I believe that Turing's own consciousness had a lot to do with what went on (physically) in his brain, and I believe there's no way getting around that point.
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
|
Jun 24, 2017 - 08:51am PT
|
I don't think you can have it both ways. I posted way upthread some ideas from Feyerabend which very clearly spelled out the consequences of a successful scientific explanation of anything, that is, the explanation provides a means of generating the same behavior as the actual physical system.
And so it would be for "mind," that is, when a scientific explanation of mind becomes available, it will be possible to simulate the mind apart from the body. A very successful theory would produce a non-biological mind indistinguishable from the real deal. This is also the insight behind the "Turing Test."
If you assign some "innate" property to the brain/body that produces mind, you immediately run into problems. Largo and others posit that "consciousness" is a physical attribute that exists as a property of matter (?!). Unfortunately there doesn't seem to be any other phenomenon than consciousness that requires this physical attribute, and not only that, but since there is no definition of "consciousness" it isn't very clear just what this attribute does.
We could dismiss this criticism by saying the science of the mind doesn't conform to all other science.
I don't think it is a productive avenue of research, and the track record of this approach confirms my suspicions. The extended several millennia written record of this sort of discussion has failed to come to a conclusion regarding the OP question.
So as Largo pointedly argues, if you believe that "mind" cannot be separated from body, then you have given up on a scientific explanation of mind all together.
This issue is so contentious that the largest attempt to externalize mind, and to use that research to predict its behaviors has largely gone unnoticed, the commercial uses of "Big Data" are precisely that.
What is more important to the "free market" than the "freedom to make choices," aka your free will. So how is it that companies, like Amazon and Google, have made it a business to watch your buying behavior, and are able to recommend products to you that you might buy? And they do it in a very directed way.
One of my friends received a book from Amazon in the mail, "...we thought you'd like this, if you don't send it back to us...". Big Data attempts to create an entirely empirically based model of your behavior, and use it predictively, to sell you stuff. It side steps the thorny philosophical issues by just ignoring them, and believes that some massive linear solver might just be the answer... ok, maybe it is an iterative linear solver.
We can laugh at such an enterprise as being ridiculously simple minded and naive, but it has had its successes, and these are not trivial when looked at from the underlying basis of our buying behaviors, which are firmly routed in our evolutionary development.
It wasn't so long ago that resources were scarce and there was a huge value in getting the most out of every resource we acquired. Killing a large animal resulted in good eats, but little of that animal didn't wind up as some other product, skins for clothing, shelter, bedding, etc, bone for tools, sinew for binding.
The point is our behavior towards resources, their acquisition and use, are well developed and a part of our species' genetic inheritance. And those behaviors are being "hacked" by large internet firms that have access at a very high level to those very behaviors, as evidenced by our online activities.
And they are very successful at it.
How is that possible? Unless our notions of free-will are not quite what we claim they are, and the touted "mystery of mind" not as mysterious as art would claim, maybe it is really banal, commercial.
|
|
yanqui
climber
Balcarce, Argentina
|
|
|
Jun 24, 2017 - 08:57am PT
|
a successful scientific explanation
Where did I say I believed that was possible?
the explanation provides a means of generating the same behavior as the actual physical system.
Ok, this sounds kind of reasonable. When the robotic guys build a robot that can autonomously have a successful career as a professional baseball player or or a profesional mathematician (say can successfully engage in the task autonomously for a 10 year period, interacting with peers and performing all the duties) I'll have to say that's some convincing evidence of a conscious machine.
Watch out if that happens!
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
|
Jun 24, 2017 - 09:15am PT
|
the professional robotic mathematician will happen before the basketball player... and at least the mathematicians will embrace the idea, NBA fans? not so much.
I didn't say you believed there was a scientific explanation for mind... I was riffing off your post.
|
|
|
SuperTopo on the Web
|
Let us know!