Re: [agi] Learning without Understanding?
The only thing I find surprising in that story is: "The findings go against one prominent theory that says children can only show smart, flexible behavior if they have conceptual knowledge – knowledge about how things work..." I don't see how anybody who's watched human beings at all can come with such a theory. People -- not just children -- do so much by rote, "because that's the way we do things here", come up with totally clueless scientific theories like this, and so forth. Joe and Bob are carpenters, working on a house. Joe is hammering and Bob is handing him the nails. Bob says, "Hey, wait a minute, half of these nails are defective." He takes out a nail and holds it up and sure enough, the head is toward the wall and the point is toward the hammer. Joe retorts, "Those aren't defective, you idiot, they're for the other side of the house." Josh --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] Roadrunner PetaVision
Derek Zahn wrote: Brain modeling certainly does seem to be in the news lately. Checking out nextbigfuture.com, I was reading about that petaflop computer Roadrunner and articles about it say that they are or will soon be emulating the entire visual cortex -- a billion neurons. I'm sure I'm not the only one who thinks that knowing what the cortex does and roughly how it does it could be quite inspiring for AGI, so I was surprised at this news. Does anybody have links to more information (besides the short recent mainstream news story)? Are they just being enthusiastic about their big computer or do they have a sophisticated theory? I don't have more information, but I would counsel caution. In my past experiences with claims of this sort (i.e. "they are or will soon be emulating the entire visual cortex") it turns out that when you ask for the exact details of the project, you find that "entire" visual cortex means something like "we are going to sample one neuron in 10,000 and measure 10% of its connections, then extrapolate from this to an entire visual cortex". Unless someone can convince me that they are going to scan a complete visual cortex in such detail that they can track all connections, right down to the individual synaptic boutons, and then translate that into a precise computational model that takes account of all the molecular mechanisms that play some role in signal transmission, I am not going to buy it. In fact, claims like this have become so outrageously exaggerated that, these days, I cannot even be bothered to move the mouse far enough to click on a link and go find out what the real story is. Their big computer might be able to model *something*, but it sounds like marketing hype to call that something "the entire visual cortex". Richard Loosemore --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] Learning without Understanding?
Brad Paulsen wrote: Hear Ye, Hear Ye... CHILDREN LEARN SMART BEHAVIORS WITHOUT KNOWING WHAT THEY KNOW http://www.physorg.com/news132839991.html It's garbage science. Or at least, it is a garbage headline. There is a whole body of experiments done with adults in which subjects are asked to learn about several conceptual categories as a result of seeing only exemplars of the categories, without ever being told explicitly what the reasons are for a given instance being in one category or another. These adults can easily pick up the categories even when they cannot easily articulate what the criteria are. This is concept building, and it is one of the most fundamental activities of the human mind. Is it surprising or new that children do the same thing? It should be stupidly obvious that they do the same thing. Children spend all their time voraciously separating the world out into categories, using almost nothing but exemplar-based learning. Just because I believe that there is much of value in cognitive science, doesn't mean I will defend everything done in its name. Richard Loosemore --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
Mike A.: Well, if you're convinced that infinity and the uncomputable are imaginary things, then you've got a self-consistent view that I can't directly argue against. But are you really willing to say that seemingly understandable notions such as the problem of deciding whether a given Turing machine will eventually halt are nonsense, simply because we would need infinite time to verify that one doesn't halt? Ben J.: "Step 3 requires human society to invent new concepts and techniques, and to thereby perform hypercomputation. I don't think that a computable nonmonotonic logic really solves this problem." I agree that nonmonotonic logic is not enough, not nearly. The point is just that since there are computable approximations of hypercomputers, it is not unreasonable to allow an AGI to reason about uncomputable objects. "My own interpretation of the work is that an individual person is no more powerful than a Turing machine (though, this point isn't discussed in the paper), but that society as a whole is capable of hypercomputation because we can keep drawing upon more resources to solve a problem: we build machines, we reproduce, we interact with and record our thoughts in the environment. Effectively, society as a whole becomes somewhat like a Zeus machine - faster and more complex with each moment." Something like this is mentioned in the paper as objection #4. But personally, I'd respond as follows: if a society of AGIs can hypercompute, then why not a single AGI with a society-of-mind style architecture? It is difficult to distinguish between a closely-linked society and a loosely-knit individual, where AI is concerned. So I argue that if a society can (and should) hypercompute, there is no reason to suspect that an individual can't (or shouldn't). On Mon, Jun 16, 2008 at 11:37 PM, Mike Archbold <[EMAIL PROTECTED]> wrote: >> I'm not sure that I'm responding to your intended meaning, but: all >> computers are in reality finite-state machines, including the brain >> (granted we don't think the real-number calculations on the cellular >> level are fundamental to intelligence). However, the finite state >> machines we call PCs are so large that it is convenient to pretend >> they have infinite memory; and when we do this, we get a machine that >> is equivalent in power to a Turing machine. But a turing machine has >> an infinite tape, so it cannot really exist (the real computer >> eventually runs out of memory). Similarly, I'm arguing that the human >> brain is so large in particular ways that it is convenient to treat it >> as an even more powerful machine (perhaps an infinite-time turing >> machine), despite the fact that such a machine cannot exist (we only >> have a finite amount of time to think). Thus a "spurious infinity" is >> not so spurious. > > Abrahm, > > Thanks for responding. You know, i might be in a bit over my head with > some of the terminology in your paper, so to apologize in advance, but > just to clarify: "spurious infinity" according to Hegel is the sleight of > hand the happens when quantity transitions surreptiously into a quality. > At some point counting up, we are simply not talking about any number at > all, but about a quality of being REALLY SUPER BIG as we make kind of a > leap. > > According to him when we talk about infinity we are talking about some > idea of a huge number (in this case of calculations) and to use a phrase > he liked: "imaginary being." So since I am kind of a Hegelian of sorts > when I scanned the paper it looked like it argued that it is not possible > to compute something that I had become convinced was imaginary anyway. > That would be true if you bought into Hegel's definition of infinity and I > realize there aren't a log of hegelians around. But, tomorrow I will read > further. > > Mike > > >> >> On Mon, Jun 16, 2008 at 9:19 PM, Mike Archbold <[EMAIL PROTECTED]> wrote: I previously posted here claiming that the human mind (and therefore an >>> ideal AGI) entertains uncomputable models, counter to the AIXI/Solomonoff model. There was little enthusiasm about this idea. :) >>> Anyway, I hope I'm not being too annoying if I try to argue the point >>> once again. This paper also argues the point: http://www.osl.iu.edu/~kyross/pub/new-godelian.pdf >>> >>> It looks like the paper hinges on: >>> "None of this prior work takes account of G¨odel intuition, repeatedly >>> communicated >>> to Hao Wang, that human minds "converge to infinity" in their power, and >>> for this reason >>> surpass the reach of ordinary Turing machines." >>> >>> The thing to watch out for here is what Hegel described as the "spurious >>> infinity" which is just the imagination thinking some imaginary quantity >>> really big, but no matter how big, you always can envision "+1", but the >>> result is always just another imaginary big number, to which you can add >>> another "+1"... the point being that infinity is a idealistic quality, >>> not >
Re: [agi] the uncomputable
On Tue, Jun 17, 2008 at 9:10 PM, Abram Demski <[EMAIL PROTECTED]> wrote: > Mike A.: > > Well, if you're convinced that infinity and the uncomputable are > imaginary things, then you've got a self-consistent view that I can't > directly argue against. But are you really willing to say that > seemingly understandable notions such as the problem of deciding > whether a given Turing machine will eventually halt are nonsense, > simply because we would need infinite time to verify that one doesn't > halt? > Every thing that you understand is "imaginary", your understanding itself is an image in your mind, which could get there reflecting reality, through limited number of steps (or so physicists keep telling), or could be generated by overly vivid finite imagination. No nonsense, just finite sense. What is this with verification that a machine doesn't halt? One can't do it, so what is the problem? -- Vladimir Nesov [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
"No nonsense, just finite sense. What is this with verification that a machine doesn't halt? One can't do it, so what is the problem?" The idea would be (if Mike is really willing to go that far): "It makes sense to say that a given Turing machine DOES halt; I know what that means. But to say that one DOESN'T halt? How can I make sense of that? Either a given machine has halted, or it has not halted yet. But to say that it never halts requires infinity, a nonsensical concept." An AI that only understood computable concepts would agree with the above. What I am saying is that such a view is... inhuman. On Tue, Jun 17, 2008 at 1:29 PM, Vladimir Nesov <[EMAIL PROTECTED]> wrote: > On Tue, Jun 17, 2008 at 9:10 PM, Abram Demski <[EMAIL PROTECTED]> wrote: >> Mike A.: >> >> Well, if you're convinced that infinity and the uncomputable are >> imaginary things, then you've got a self-consistent view that I can't >> directly argue against. But are you really willing to say that >> seemingly understandable notions such as the problem of deciding >> whether a given Turing machine will eventually halt are nonsense, >> simply because we would need infinite time to verify that one doesn't >> halt? >> > > Every thing that you understand is "imaginary", your understanding > itself is an image in your mind, which could get there reflecting > reality, through limited number of steps (or so physicists keep > telling), or could be generated by overly vivid finite imagination. > > No nonsense, just finite sense. What is this with verification that a > machine doesn't halt? One can't do it, so what is the problem? > > -- > Vladimir Nesov > [EMAIL PROTECTED] > > > --- > agi > Archives: http://www.listbox.com/member/archive/303/=now > RSS Feed: http://www.listbox.com/member/archive/rss/303/ > Modify Your Subscription: http://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com > --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] Learning without Understanding?
--- On Tue, 6/17/08, Brad Paulsen <[EMAIL PROTECTED]> wrote: > CHILDREN LEARN SMART BEHAVIORS WITHOUT KNOWING WHAT THEY KNOW > http://www.physorg.com/news132839991.html Another example: children learn to form grammatically correct sentences before they understand the difference between a noun and a verb. -- Matt Mahoney, [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
On Tue, Jun 17, 2008 at 10:14 PM, Abram Demski <[EMAIL PROTECTED]> wrote: > "No nonsense, just finite sense. What is this with verification that a > machine doesn't halt? One can't do it, so what is the problem?" > > The idea would be (if Mike is really willing to go that far): "It > makes sense to say that a given Turing machine DOES halt; I know what > that means. But to say that one DOESN'T halt? How can I make sense of > that? Either a given machine has halted, or it has not halted yet. But > to say that it never halts requires infinity, a nonsensical concept." > > An AI that only understood computable concepts would agree with the > above. What I am saying is that such a view is... inhuman. > It wasn't worded correctly, there are many machines that you can prove don't halt, but also others for which you can't prove that. Why would that be inhuman to not be able to do impossible? -- Vladimir Nesov [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
> Mike A.: > > Well, if you're convinced that infinity and the uncomputable are > imaginary things, then you've got a self-consistent view that I can't > directly argue against. But are you really willing to say that > seemingly understandable notions such as the problem of deciding > whether a given Turing machine will eventually halt are nonsense, > simply because we would need infinite time to verify that one doesn't > halt? > Abrahm, I don't think I really disagree with you, and like I said, I need to go back and really read the whole paper. It looks interesting. I will make a comment that is not a counterargument but rather something to think about. By imaginary I mean that it cannot really be pointed at. I was thinking about this... you know, if we took a rule base, a finite set of rules, and let's say we start with that -- clearly finite. We can identify each rule and literally point to the rule on the screen (I guess, think of Cyc). We have a QUANTITY of rules. Then suppose we have determined that in order to simulate human thought, we must make the rule base infinite, as our mathematical/philosophical investigations have shown that to achieve AGI the infinite is necessary, let's say infinite time and quantity. At that point the finite quantity of rules would disappear and we would cross over strictly into the imaginary idea of a really huge number of rules and a huge imaginary time: that is, we are not speaking of a QUANTITY of rules and time -- we have expressly and surreptitiously shifted from discussing quantities to qualities. Whereas before we were talking about a finite set of rules, now we fold our hands and say, "well, it's an infinite set of rules and time now." We have shifted from a finite-quantity that could be examined to an infinity-quality that cannot be examined and is wholly imaginary, yet using that as a proof. When we make this transition, it seems to me that the shift is so radical that it is impossible to justify making the step, because as I mentioned it involves a surreptitious shift from quantity to quality. Incidentally Hegel held that the "true infinite" (as opposed to the spurious infinite which is the unwarranted transition from quantity to quality) was human thought. I've been working on a book written by David Carlson, a law professor, which makes clear some of the very obscure writing of Hegel. Mike > Ben J.: > > "Step 3 requires human society to invent new concepts and techniques, > and to thereby perform hypercomputation. I don't think that a > computable nonmonotonic logic really solves this problem." > > I agree that nonmonotonic logic is not enough, not nearly. The point > is just that since there are computable approximations of > hypercomputers, it is not unreasonable to allow an AGI to reason about > uncomputable objects. > > "My own interpretation of the work is that an individual person is no more > powerful than a Turing machine (though, this point isn't discussed in the > paper), but that society as a whole is capable of hypercomputation because > we can keep drawing upon more resources to solve a problem: we build > machines, we reproduce, we interact with and record our thoughts in the > environment. Effectively, society as a whole becomes somewhat like a Zeus > machine - faster and more complex with each moment." > > Something like this is mentioned in the paper as objection #4. But > personally, I'd respond as follows: if a society of AGIs can > hypercompute, then why not a single AGI with a society-of-mind style > architecture? It is difficult to distinguish between a closely-linked > society and a loosely-knit individual, where AI is concerned. So I > argue that if a society can (and should) hypercompute, there is no > reason to suspect that an individual can't (or shouldn't). > > On Mon, Jun 16, 2008 at 11:37 PM, Mike Archbold <[EMAIL PROTECTED]> > wrote: >>> I'm not sure that I'm responding to your intended meaning, but: all >>> computers are in reality finite-state machines, including the brain >>> (granted we don't think the real-number calculations on the cellular >>> level are fundamental to intelligence). However, the finite state >>> machines we call PCs are so large that it is convenient to pretend >>> they have infinite memory; and when we do this, we get a machine that >>> is equivalent in power to a Turing machine. But a turing machine has >>> an infinite tape, so it cannot really exist (the real computer >>> eventually runs out of memory). Similarly, I'm arguing that the human >>> brain is so large in particular ways that it is convenient to treat it >>> as an even more powerful machine (perhaps an infinite-time turing >>> machine), despite the fact that such a machine cannot exist (we only >>> have a finite amount of time to think). Thus a "spurious infinity" is >>> not so spurious. >> >> Abrahm, >> >> Thanks for responding. You know, i might be in a bit over my head with >> some of the terminology in your paper, so
Re: [agi] the uncomputable
V. N., What is inhuman to me, is to claim that the halting problem is no problem on such a basis: that the statement "Turing machine X does not halt" only is true of Turing machines that are *provably* non-halting. And this is the view we are forced into if we abandon the reality of the uncomputable. A. D. On Tue, Jun 17, 2008 at 2:34 PM, Vladimir Nesov <[EMAIL PROTECTED]> wrote: > On Tue, Jun 17, 2008 at 10:14 PM, Abram Demski <[EMAIL PROTECTED]> wrote: >> "No nonsense, just finite sense. What is this with verification that a >> machine doesn't halt? One can't do it, so what is the problem?" >> >> The idea would be (if Mike is really willing to go that far): "It >> makes sense to say that a given Turing machine DOES halt; I know what >> that means. But to say that one DOESN'T halt? How can I make sense of >> that? Either a given machine has halted, or it has not halted yet. But >> to say that it never halts requires infinity, a nonsensical concept." >> >> An AI that only understood computable concepts would agree with the >> above. What I am saying is that such a view is... inhuman. >> > > It wasn't worded correctly, there are many machines that you can prove > don't halt, but also others for which you can't prove that. Why would > that be inhuman to not be able to do impossible? > > -- > Vladimir Nesov > [EMAIL PROTECTED] > > > --- > agi > Archives: http://www.listbox.com/member/archive/303/=now > RSS Feed: http://www.listbox.com/member/archive/rss/303/ > Modify Your Subscription: http://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com > --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
People interested on this thread subject might be interested to read a paper we wrote some years ago published by World Scientific: --- Hector Zenil, Francisco Hernandez-Quiroz, "On the possible Computational Power of the Human Mind", WORLDVIEWS, SCIENCE AND US, edited by Carlos Gershenson, Diederik Aerts and Bruce Edmonds, World Scientific, 2007. available online: http://arxiv.org/abs/cs/0605065 Abstract The aim of this paper is to address the question: Can an artificial neural network (ANN) model be used as a possible characterization of the power of the human mind? We will discuss what might be the relationship between such a model and its natural counterpart. A possible characterization of the different power capabilities of the mind is suggested in terms of the information contained (in its computational complexity) or achievable by it. Such characterization takes advantage of recent results based on natural neural networks (NNN) and the computational power of arbitrary artificial neural networks (ANN). The possible acceptance of neural networks as the model of the human mind's operation makes the aforementioned quite relevant. Presented as a talk at the Complexity, Science and Society Conference, 2005, University of Liverpool, UK. --- On the other hand, Goedelian type arguments (such as http://www.osl.iu.edu/~kyross/pub/new-godelian.pdf) have been widely accepted to be disproved since Hofstadter's Escher, Goedel and Bach in the 70s or before. I consider myself as someone within the busy beaver field since my own research on what we call experimental algorithmic information theory is very related to. I don't see how either Solomonoff's induction or the Busy Beaver problem can be used as evidence or be conceived as an explaination of the human mind as a hypercomputer. I don't see in the development of the two fields anything not Turing computable. There are known values of the busy beaver up to 4 state 2 symbol Turing machines (although it seems they claim to have calculated up to 6 states...). To determine whether a Turing machine halts up to that number of states is a relatively easy task by using very computable tricks, (including the Christmas Tree method). I think their main argument is that (a) once known the value of a busy beaver for n states, one learns how to crack the set of n+1 states and eventually get it. (i) They then use a kind of mathematical induction to proof that any given Turing machine with a fixed number of states will eventually fail, while the human mind can go on. However it seems pretty clear that the method evidently fails for n large enough, and hence disproving their claim. Now suppose their claim is right (a), now let's conceive the following method: (b) that each time we learn how to crack n+1 we build a Turing machine T that computes n+1, using their own argument (i) then Turing machine are hypercomputers! I might be missing something, if so please feel free to point it out. Best regards, -- Hector Zenilhttp://zenil.mathrix.org --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
On Tue, Jun 17, 2008 at 11:38 PM, Abram Demski <[EMAIL PROTECTED]> wrote: > V. N., > What is inhuman to me, is to claim that the halting problem is no > problem on such a basis: that the statement "Turing machine X does not > halt" only is true of Turing machines that are *provably* non-halting. > And this is the view we are forced into if we abandon the reality of > the uncomputable. > Why, you can also mark up the remaining territory by "true" and "false", these labels just won't mean anything there. Set up to sets, T and F, place all true things in T, all false things in F, and all unknown things however you like, but don't tell anybody how. Some people like to place all unknown things in F, their call. Mathematically it can be convenient, but really, even of "computable" things you can't really compute that much, so the argument is void for all practical concerns anyway. -- Vladimir Nesov [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
Vladimir Nesov, Then do you agree with my hypothetical extremist version of Mike? (Aside: For the example we are talking about, it is totally necessary to stick the undecidable cases in F rather than T: if a Turing machine halts, then it is possible to prove that it halts (simply by running it for long enough). So if a Turing machine is one of those whose halting is formally undecidable, then it must not halt, because if it did then a proof of its halting would exist.) Hector Zenil, I do not think I understand you. Your argument seems similar to the following: "I do not see why Turing machines are necessary. If we can compute a function f(x) by some Turing machine, then we could compute it up to some value x=n. But we could construct a lookup table of all values f(0), f(1), f(2),... , f(n) which contains just as much information." Obviously the above is a silly argument, but I don't know how else to interpret you. A Turing machine can capture a finite number of the outputs of a hypercomputer. Does that in any way make the hypercomputer reducible to the Turing machine? Mike Archbold, It seems you've made a counterargument without meaning to. "When we make this transition, it seems to me that the shift is so radical that it is impossible to justify making the step, because as I mentioned it involves a surreptitious shift from quantity to quality." I maintain that the jump is justified. To me it is like observing the sequence "1, 2, 4, 8, 16, 32..." and concluding that each number is twice the previous. It is a jump from several quantities to a single quality. On Tue, Jun 17, 2008 at 4:35 PM, Vladimir Nesov <[EMAIL PROTECTED]> wrote: > On Tue, Jun 17, 2008 at 11:38 PM, Abram Demski <[EMAIL PROTECTED]> wrote: >> V. N., >> What is inhuman to me, is to claim that the halting problem is no >> problem on such a basis: that the statement "Turing machine X does not >> halt" only is true of Turing machines that are *provably* non-halting. >> And this is the view we are forced into if we abandon the reality of >> the uncomputable. >> > > Why, you can also mark up the remaining territory by "true" and > "false", these labels just won't mean anything there. Set up to sets, > T and F, place all true things in T, all false things in F, and all > unknown things however you like, but don't tell anybody how. Some > people like to place all unknown things in F, their call. > Mathematically it can be convenient, but really, even of "computable" > things you can't really compute that much, so the argument is void for > all practical concerns anyway. > > -- > Vladimir Nesov > [EMAIL PROTECTED] > > > --- > agi > Archives: http://www.listbox.com/member/archive/303/=now > RSS Feed: http://www.listbox.com/member/archive/rss/303/ > Modify Your Subscription: http://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com > --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
On Tue, Jun 17, 2008 at 5:58 PM, Abram Demski <[EMAIL PROTECTED]> wrote: > > Hector Zenil, > > I do not think I understand you. Your argument seems similar to the following: > > "I do not see why Turing machines are necessary. If we can compute a > function f(x) by some Turing machine, then we could compute it up to > some value x=n. But we could construct a lookup table of all values > f(0), f(1), f(2),... , f(n) which contains just as much information." > > Obviously the above is a silly argument, but I don't know how else to > interpret you. A Turing machine can capture a finite number of the > outputs of a hypercomputer. Does that in any way make the > hypercomputer reducible to the Turing machine? > This nicely boils the fallacy down from 20 pages to a few lines. Merely providing the lookup table or adding more states is not sufficient to turn a Turing machine into a hypercomputer as it would follow from the paper main argument: that humans can always find bb(n+1) once bb(n) calculated, therefore humans are capable of hypercomputing (modulo other strong assumptions). In fact the paper acknowledges that more information is needed at each jump, so eventually one would reach either a physical or a feasible limit unless the brain/mind is infinite in capabilities, falling into the traditional claims on hypercomputation, and not necessarily a new one. I recall that my suggestion was (reductio ad absurdum) to encode (or provide the program) a n-state Turing machine T_n after knowing bb(n) so at every moment when people is working on bb(n+1) there is always a T_n behind able to calculate bb(n). Once the hyperhuman finds bb(n+1) then he encodes T_{n+1} to compute bb(n+1) while the hyperhuman H computes bb(n+2) but one knows that at the next step one will be able to code T_{n+2} to calculate bb(n+2), just as H does. Following their argument, if there is always a machine able to calculate bb(n+1) for any n when bb(n) is calculated (as there is a hyperhuman according to their claim), therefore T (the universal Turing machine that emulates all those T_i for all i) would turn into a hypercomputer (absurd since it would collapse the classes of computability!). Notice that my use of hypercomputer is the traditional use of a computer: a machine able to compute at a Turing degree other than the first. I still might be missing something, but hope this clarifies my objection. People might be also interested in the work of Kevin Kelly: "Uncomputability: The Problem of Induction Internalized," Theoretical Computer Science, pp. 317: 2004, 227-249. as an epistemological approach to traditional computability, as some have suggested in this thread induction as evidence for hypercomputability. -- Hector Zenilhttp://zenil.mathrix.org > On Tue, Jun 17, 2008 at 4:35 PM, Vladimir Nesov <[EMAIL PROTECTED]> wrote: >> On Tue, Jun 17, 2008 at 11:38 PM, Abram Demski <[EMAIL PROTECTED]> wrote: >>> V. N., >>> What is inhuman to me, is to claim that the halting problem is no >>> problem on such a basis: that the statement "Turing machine X does not >>> halt" only is true of Turing machines that are *provably* non-halting. >>> And this is the view we are forced into if we abandon the reality of >>> the uncomputable. >>> >> >> Why, you can also mark up the remaining territory by "true" and >> "false", these labels just won't mean anything there. Set up to sets, >> T and F, place all true things in T, all false things in F, and all >> unknown things however you like, but don't tell anybody how. Some >> people like to place all unknown things in F, their call. >> Mathematically it can be convenient, but really, even of "computable" >> things you can't really compute that much, so the argument is void for >> all practical concerns anyway. >> >> -- >> Vladimir Nesov >> [EMAIL PROTECTED] >> >> >> --- >> agi >> Archives: http://www.listbox.com/member/archive/303/=now >> RSS Feed: http://www.listbox.com/member/archive/rss/303/ >> Modify Your Subscription: http://www.listbox.com/member/?&; >> Powered by Listbox: http://www.listbox.com >> > > > --- > agi > Archives: http://www.listbox.com/member/archive/303/=now > RSS Feed: http://www.listbox.com/member/archive/rss/303/ > Modify Your Subscription: http://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com > --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
> Mike Archbold, > > It seems you've made a counterargument without meaning to. > > "When we make this transition, it seems to me that the shift is so radical > that it is impossible to justify making the step, because as I mentioned > it involves a surreptitious shift from quantity to quality." > > I maintain that the jump is justified. To me it is like observing the > sequence "1, 2, 4, 8, 16, 32..." and concluding that each number is > twice the previous. It is a jump from several quantities to a single > quality. > Fair enough. I think what we are saying is that in the transition of quantity to quality -- as in your example -- is a kind of "appeal to the infinite," ie., is an instance of hypercompuation? That does have a Godel sound to it, as I understand it, we appeal beyond the data in question although I have only seen Godel's writings (not read them). I read more of the paper. I like the part about the Zeus Machine. Cool. I guess I am a bit more aligned to the philosophy side than the Turing-Godel-computational side of the house. In my studies as I mentioned of Hegel's Logic, there is a constant interplay between quantity and quality, given as measure -- measure here being the result of quantity and quality intermixing. I guess measure in this sense is roughly equivalent to hypercomputation if I have my Godels and Hegels lined up in a row. Hegel's philosophy was of course totally predicated on the mind which as I said he held to be infinite. Although, we have to be careful in so much as there exist multiple definitions of the infinite. Mike Archbold > On Tue, Jun 17, 2008 at 4:35 PM, Vladimir Nesov <[EMAIL PROTECTED]> > wrote: >> On Tue, Jun 17, 2008 at 11:38 PM, Abram Demski <[EMAIL PROTECTED]> >> wrote: >>> V. N., >>> What is inhuman to me, is to claim that the halting problem is no >>> problem on such a basis: that the statement "Turing machine X does not >>> halt" only is true of Turing machines that are *provably* non-halting. >>> And this is the view we are forced into if we abandon the reality of >>> the uncomputable. >>> >> >> Why, you can also mark up the remaining territory by "true" and >> "false", these labels just won't mean anything there. Set up to sets, >> T and F, place all true things in T, all false things in F, and all >> unknown things however you like, but don't tell anybody how. Some >> people like to place all unknown things in F, their call. >> Mathematically it can be convenient, but really, even of "computable" >> things you can't really compute that much, so the argument is void for >> all practical concerns anyway. >> >> -- >> Vladimir Nesov >> [EMAIL PROTECTED] >> >> >> --- >> agi >> Archives: http://www.listbox.com/member/archive/303/=now >> RSS Feed: http://www.listbox.com/member/archive/rss/303/ >> Modify Your Subscription: http://www.listbox.com/member/?&; >> Powered by Listbox: http://www.listbox.com >> > > > --- > agi > Archives: http://www.listbox.com/member/archive/303/=now > RSS Feed: http://www.listbox.com/member/archive/rss/303/ > Modify Your Subscription: > http://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com > --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] Learning without Understanding?
- Original Message From: Richard Loosemore <[EMAIL PROTECTED]> Brad Paulsen wrote: > CHILDREN LEARN SMART BEHAVIORS WITHOUT KNOWING WHAT THEY KNOW > http://www.physorg.com/news132839991.html It's garbage science. Or at least, it is a garbage headline. There is a whole body of experiments done with adults in which subjects are asked to learn about several conceptual categories as a result of seeing only exemplars of the categories, without ever being told explicitly what the reasons are for a given instance being in one category or another. These adults can easily pick up the categories even when they cannot easily articulate what the criteria are. This is concept building, and it is one of the most fundamental activities of the human mind. Is it surprising or new that children do the same thing? It should be stupidly obvious that they do the same thing. Children spend all their time voraciously separating the world out into categories, using almost nothing but exemplar-based learning. Just because I believe that there is much of value in cognitive science, doesn't mean I will defend everything done in its name. Richard Loosemore -- Well, cognitive science progresses by questioning other conclusions and then devising new experiments that can produce more insightful results. One of the problems with this kind of experiment is that children in the (relatively) more affluent communities of the industrialized world already have a (relatively) sophisticated capability to assess certain aspects of images on a video screen. The fact that a group of cognitive scientists might be totally unaware of the potential significance of this kind of complex awareness is an oopsie that can only be due to the innocence of youth. I wonder what the average age of the researchers were and if they fully realized what they were doing? But the issue so important that the experiment does deserve some attention. If a more sophisticated set of experiments could provide more detail about how implicit knowledge is acquired and becomes explicit, then the results might be important. Jim Bromer --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
[agi] Have you hugged a cephalopod today?
From "The More stuff we already know" department... NEW RESEARCH ON OCTOPUSES SHEDS LIGHT ON MEMORY http://www.physorg.com/news132920831.html Cheers, Brad --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
Re: [agi] the uncomputable
On Wed, Jun 18, 2008 at 1:58 AM, Abram Demski <[EMAIL PROTECTED]> wrote: > Vladimir Nesov, > > Then do you agree with my hypothetical extremist version of Mike? > > (Aside: For the example we are talking about, it is totally necessary > to stick the undecidable cases in F rather than T: if a Turing machine > halts, then it is possible to prove that it halts (simply by running > it for long enough). So if a Turing machine is one of those whose > halting is formally undecidable, then it must not halt, because if it > did then a proof of its halting would exist.) > If T* is enumerable set of halting machines, F* is enumerable (by our machine) subset of never-halting machines, and X is set of remaining machines, the situation is symmetric. If you dump X into T*, you destroy its property to be enumerable, but likewise with F*. -- Vladimir Nesov [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com