Re: [agi] Goal Driven Systems and AI Dangers [WAS Re: Singularity Outcomes...]
Steve Richfield wrote: ... have played tournament chess. However, when faced with a REALLY GREAT chess player (e.g. national champion), as I have had the pleasure of on a couple of occasions, they at first appear to play as novices, making unusual and apparently stupid moves that I can't quite capitalize on, only to pull things together later on and soundly beat me. While retrospective analysis would show them to be brilliant, that would not be my evaluation early in these games. Steve Richfield But that's a quite reasonable action on their part. Many players have memorized some number of standard openings. But by taking the game away from the standard openings (or into the less commonly known ones) they enable the player with the stronger chess intuition to gain an edge...and they believe that it will be themselves. E.g.: The Orangutan opening is a trifle weak, but few know it well. But every master would know it, and know both it's strengths and weaknesses. If you don't know the opening, though, it just looks weak. Looks, however, are deceptive. If you don't know it, you're quite likely to find it difficult to deal with against someone who does know it, even if they're a generally weaker player than you are. --- 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: AW: [agi] How general can be and should be AGI?
Mike Tintner wrote: Charles: Flaws in Hamlet: I don't think of this as involving general intelligence. Specialized intelligence, yes, but if you see general intelligence at work there you'll need to be more explicit for me to understand what you mean. Now determining whether a particular deviation from iambic pentameter was a flaw would require a deep human intelligence, but I don't feel that understanding of how human emotions are structured is a part of general intelligence except on a very strongly superhuman level. The level where the AI's theory of your mind was on a par with, or better than, your own. Charles, My flabber is so ghasted, I don't quite know what to say. Sorry, I've never come across any remarks quite so divorced from psychological reality. There are millions of essays out there on Hamlet, each one of them different. Why don't you look at a few?: http://www.123helpme.com/search.asp?text=hamlet I've looked at a few (though not those). In college I formed the definite impression that essays on the meaning of literature were exercises in determining what the instructor wanted. This isn't something that I consider a part of general intelligence (except as mentioned above). ... The reason over 70 per cent of students procrastinate when writing essays like this about Hamlet, (and the other 20 odd per cent also procrastinate but don't tell the surveys), is in part that it is difficult to know which of the many available approaches to take, and which of the odd thousand lines of text to use as support, and which of innumerable critics to read. And people don't have a neat structure for essay-writing to follow. (And people are inevitably and correctly afraid that it will all take if not forever then far, far too long). The problem is that most, or at least many, of the approaches are defensible, but your grade will be determined by the taste of the instructor. This isn't a problem of general intelligence except at a moderately superhuman level. Human tastes aren't reasonable ingredients for an entry level general intelligence. Making it a requirement merely ensures that one will never be developed (whose development attends to your theories of what's required). ... In short, essay writing is an excellent example of an AGI in action - a mind freely crossing different domains to approach a given subject from many fundamentally different angles. (If any subject tends towards narrow AI, it is normal as opposed to creative maths). I can see story construction as a reasonable goal for an AGI, but at the entry level they are going to need to be extremely simple stories. Remember that the goal structures of the AI won't match yours, so only places where the overlap is maximal are reasonable grounds for story construction. Otherwise this is an area for specialized AIs, which isn't what we are after. Essay writing also epitomises the NORMAL operation of the human mind. When was the last time you tried to - or succeeded in concentrating for any length of time? I have frequently written essays and other similar works. My goal structures, however, are not generalized, but rather are human. I have built into me many special purpose functions for dealing with things like plot structure, family relationships, relative stages of growth, etc. As William James wrote of the normal stream of consciousness: Instead of thoughts of concrete things patiently following one another in a beaten track of habitual suggestion, we have the most abrupt cross-cuts and transitions from one idea to another, the most rarefied abstractions and discriminations, the most unheard-of combinations of elements, the subtlest associations of analogy; in a word, we seem suddenly introduced into a seething caldron of ideas, where everything is fizzling and bobbing about in a state of bewildering activity, where partnerships can be joined or loosened in an instant, treadmill routine is unknown, and the unexpected seems the only law. Ditto: The normal condition of the mind is one of informational disorder: random thoughts chase one another instead of lining up in logical causal sequences. Mihaly Csikszentmihalyi Ditto the Dhammapada, Hard to control, unstable is the mind, ever in quest of delight, When you have a mechanical mind that can a) write essays or tell stories or hold conversations [which all present the same basic difficulties] and b) has a fraction of the difficulty concentrating that the brain does and therefore c) a fraction of the flexibility in crossing domains, then you might have something that actually is an AGI. You seem to be placing an extremely high bar in place before you will consider something an AGI. Accepting all that you have said, for an AGI to react as a human would react would require that the AGI be strongly superhuman. More to the point, I wouldn't DARE create an AGI which had motivations similar to
Re: AW: [agi] How general can be and should be AGI?
Dr. Matthias Heger wrote: Performance not an unimportant question. I assume that AGI has necessarily has costs which grow exponentially with the number of states and actions so that AGI will always be interesting only for toy domains. My assumption is that human intelligence is not truly general intelligence and therefore cannot hold as a proof of existence that AGI is possible. Perhaps we see more intelligence than there really is. Perhaps the human intelligence is to some extend overestimated and an illusion as the free will. Why? In truly general domains every experience of an agent only can be used for the single certain state and action when the experience was made. Every time when your algorithm makes generalizations from known state-action pairs to unknown state-action pairs then this is in fact usage of knowledge about the underlying state-action space or it is just guessing and only a matter of luck. So truly general AGI algorithms must visit every state-action pair at least once to learn what to do in what state. Even in small real world domains the state spaces are so big that it would take longer than the age of the universe to go through all states. For this reason true AGI is impossible and human intelligence must be narrow to a certain degree. I would assert a few things that appear to contradict your assumptions (and a few that suppport them). 1) AGIs will reach conclusions that are not guaranteed to be correct. This allows somewhat lossy compression of the input data. 2) AGIs can exist, but will operate in modes. In AGI mode they will be very expensive and slow. And still be error prone. 3) Humans do have an AGI mode. Probably more than one of them. But it's so expensive to use and so slow that they strive diligently to avoid using it, preferring to rely on simple situation-based models (and discarding most of the input data while doing so). 4) When humans are operating in AGI mode, they are not considering or using ANY real-time data (except to hold and replay notes). The process is too slow. The two AGI modes that I believe people use are 1) mathematics and 2) experiment. Note that both operate in restricted domains, but within those domains they *are* general. (E.g., mathematics cannot generate it's own axioms, postulates, and rules of inference, but given them it is general.) Because of the restricted domains, many problems can't even be addressed by either of them, so I suspect the presence of other AGI modes. Possibly even slower and more expensive to use. I suppose that one could quibble that since the modes I have identified are restricted to particular domains, that they aren't *general* intelligence modes, but as far as I can tell ALL modes of human thought only operate within restricted domains. --- 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: AW: [agi] How general can be and should be AGI?
Mike Tintner wrote: Charles: as far as I can tell ALL modes of human thought only operate within restricted domains. I literally can't conceive where you got this idea from :). Writing an essay - about, say, the French Revolution, future of AGI, flaws in Hamlet, what you did in the zoo, or any of the other many subject areas of the curriculum - which accounts for, at a very rough estimate, some 50% of problemsolving within education, operates within *which* restricted domain? (And how *did* you arrive at the above idea?) Yes, I think of those as being handled largely by specialized, non-general, mechanisms. I suppose that to an extent you could say that it's done via pattern matching, and to that extent it falls under the same model that I've called experimentation. Mainly, though, that's done with specialized language manipulation routines. (I'm not asserting that they are hard-wired. They were built up via lots of time and effort put in via both experimentation and mathematics [in which I include modeling and statistical prediction]). Mathematics and experimentation are extremely broad brushes. That's a part of why they are so slow. French revolution: Learning your history from a teacher or a text isn't a general pattern. It's a short-cut that usually works pretty well. Now if you were talking about going on the ground and doing personal research...then it might count as general intelligence under the category of experimentation. (Note that both mathematics and experimentation are generally necessary to creat new knowledge, rather that copying knowledge from some source that has previously acquired and processed it.) Future of AGI: Creating the future of AGI does, indeed, involve general intelligence. If you follow this list you'll note that it involves BOTH mathematics and experimentation. Flaws in Hamlet: I don't think of this as involving general intelligence. Specialized intelligence, yes, but if you see general intelligence at work there you'll need to be more explicit for me to understand what you mean. Now determining whether a particular deviation from iambic pentameter was a flaw would require a deep human intelligence, but I don't feel that understanding of how human emotions are structured is a part of general intelligence except on a very strongly superhuman level. The level where the AI's theory of your mind was on a par with, or better than, your own. --- 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] WHAT ARE THE MISSING CONCEPTUAL PIECES IN AGI?
Ed Porter wrote: WHAT ARE THE MISSING CONCEPTUAL PIECES IN AGI? One that appears to me to be missing, or at least not emphasized, is that general intelligence is inefficient compared to specialized techniques. In any particular sub-domain specialized intelligence will be able to employ better heuristics and a more appropriate and more efficient methodology. One that might well not work very well in a more general case. As a result a good AGI will have a rather large collection of specialized AIs that are a part of it's toolbox. When it encounters a new environment or problem, one of the things it will be doing, as it solves the problem of how to deal with this new problem, is build a specialized AI to handle dealing with that problem. In normal circumstances, what the AGI will do is classify the kind of problem it's dealing with, and hand it off to a more specialized AI. (And monitor the process, to make sure that the problem continues to fit the mold.) My expectation is that AGIs (without their toolbox) will be quite slow and inefficient with dealing with any particular situation. OTOH, *with* their toolbox they should be able to evolve to be more efficient than humans. (Note that efficient is relative to the hardware that they're running on. And each solution is likely to consider that a part of it's presumptions is that the hardware remains unchanged. Adapting an existing solution to new hardware is likely to be a hard problem. On a level of re-writing a program from Java into Fortran. It is my expectation that this approach will be necessary across a wide gamut of AGI designs, and that unitary minds will be scarce to non-existent. Certainly people work this way. Consider the process of learning a new musical instrument, or of learning to read music. You build a specialized module in your mind to handle the problems involved. --- 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] Comments from a lurker...
J Storrs Hall, PhD wrote: ... The third mistake is to forget that nobody knows how to program SIMD. They can't even get programmers to adopt functional programming, for god's sake; the only thing the average programmer can think in is BASIC, or C which is essentially machine-independent assembly. Not even LISP. APL, which is the closest approach to a SIMD language, died a decade or so back. ... Josh Actually I believe that Prograf (a dataflow language) had a programming model that was by far the most SIMD. Much more so than APL. It also died awhile back, trying to transition from the Mac to MSWind95. It did, however, convince me that reasonable programming idioms from SIMD were reasonable. (Actually, I think Prograf could have been implemented as MIMD. Since it was running on a single processor system, though, the actual implementation was serial.) P.S.: versions of APL still exist. The last time I checked the language was called, I believe, J. http://en.wikipedia.org/wiki/J_programming_language (Such a nice searchable name!) They eliminated the special symbols, but I don't remember what they replaced them with. Don't know if the implementation is SIMD. --- 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] Logical Satisfiability...Get used to it.
Jim Bromer wrote: Ben G wrote: ... ... Concerning beliefs and scientific rationalism: Beliefs are the basis of all thought. To imply that religious belief might be automatically different from rational beliefs is naïve. However, I think there is an advantage in defining what a rational thought is realitve to AI programming and how scientific rationalism is different from simple rationalism. I am going to write a few messages about this when I get a chance. By the way, I don't really see how a simple n^4 or n^3 SAT solver in itself would be that useful for any immediate AGI project, but the novel logical methods that the solver will reveal may be more significant. Jim Bromer But religious beliefs *ARE* intrinsically different from rational beliefs. They aren't the only such belief, but they are among them. Rational beliefs MUST be founded in other beliefs. Rationalism does not provide a basis for generating beliefs ab initio, but only via reason, which requires both axioms and rules of inference. (NARS may have eliminated the axioms, but I doubt it. OTOH, I don't understand exactly how it works yet.) Religion and other intrinsic beliefs are inherent in the construction of humans. I suspect that every intelligent entity will require such beliefs. Which particular religion is believed in isn't inherent, but is situational. (Other factors may enter in, but I would need a clear explication of how that happened before I would believe that.) Note that another inherent belief is People like me are better than people who are different. The fact that a belief is inherent doesn't mean it can't be overcome (or at least subdued) by counter-programming, merely that one will need to continually ward against it, or it will re-assert itself even if you know that it's wrong. Saying that a belief is non-rational isn't denigrating it. It's merely a statement that it isn't a built-in rule. Even the persistence of forms doesn't seem to be totally built-in, though there are definitely lots of mechanisms that will tend to create it. So in that case what's built in is a tendency to perceive the persistence of objects. In the case of religion it's a bit more difficult to perceive what the built-in process is. Plausibly it's a combination of several tendency to perceive patterns shaped like ... in the world that aren't intrinsically connected, but which have been connected by culture. Or it might be something else. (The blame/attribute everything to the big alpha baboon theory isn't totally silly, but I find it unsatisfactory. It's at most a partial answer.) --- 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] Symbols
I would suggest that symbols are more powerful than images, though less immediate (unmitigated?) in their effect. Images present a visual scene. They require processing to evaluate, and what one extracts from the scene may not be what another extracts. Their power is that they may activate a large number of symbols. Symbols, however, are different. They may contain a reference to an image, but it will be a meaningful connection. I.e., it will highlight certain features of the scene as significant, and mute the others (if they aren't elided). It will also generally contain a value (desirable-undesirable scale), and may contain a label (vocalization) and various operations that it can enter into. Think of it as being like an instance of a class in a proto-type based language. I.e., even though it is an instance, it can have derivative sub-classes. But also superclasses can freely be formed which automatically include it. In fact superclasses can be retroactively inserted into the hierarchies of symbols, and are themselves symbols. Note that most symbols *don't* have either vocalizations or visibility to consciousness. Also, symbols are a feature of the system within which they reside. They don't constitute the system. I see most of what happens as not only asymbolic, but also special purpose, such as sound recognition in the cochlea. I.e., most of what goes on is narrow AI or even hardwired. (Yes, I'm using biological metaphors. Sorry. People are the only example of something approaching an AGI that I'm aware of.) The AGI section is relatively slow and expensive, so it's kept relatively limited. Analog computation is preferred when possible, sometimes mixed with neural nets (which I think of as digital, even though I know better). Consider catching a baseball. Just try to do that with an AGI! But using the combination of analog and neural net computation that is built into people it's a reasonable task. (For most people. I'm lousy at it.) AH! But deciding to catch the baseball! That's the AGI in action. (Well, eventually it becomes habitual, I guess, for some people. But even so it's probably an AGI level decision.) Now lets consider that baseball. That's clearly a symbol, as I didn't show you a picture. You didn't know I was talking about a softball. Notice how quickly the image changed. That's because you did it by manipulating references rather than by moving around enough bits to represent an image of one or the other kind of baseball. And if I now tell you that it was bright green, the color of a tennis ball, you get yet another rapid change. This time it's probably some kind of filter being applied to the image, as I notice that even the thread holding it together immediately changed color. (In my image the thread was originally a kind of off-read, and on the green ball it shaded toward black.) But notice that this was all done via the manipulation of symbols. It depended on the images already being resident within your mind, and tied into these symbols via links (references). I haven't mentioned it, but you probably have some sort of image about the location of this playing around with the baseball. That's what I meant about ...and mute the others (if they aren't elided)..., above. (You may have totally elided the location, but all specific instances [images] that you have ever seen will have had a background.) Mark Waser wrote: Why are images almost always more powerful than the corresponding symbols? Why do they communicate so much faster? Um . . . . dude . . . . it's just a bandwidth thing. Think about images vs. visual symbols vs. word descriptions vs. names. It's a spectrum from high-bandwidth information transfer to almost pure reference tags. If it's something you've never run across before, images are best -- high bandwidth but then you end up with high mental processing costs. For familiar items, word descriptions (or better yet, single word names) require little bandwidth and little in the way of subsequent processing costs. - Original Message - *From:* Mike Tintner mailto:[EMAIL PROTECTED] *To:* agi@v2.listbox.com mailto:agi@v2.listbox.com *Sent:* Sunday, March 30, 2008 4:02 PM *Subject:* Re: [agi] Symbols In this surrounding discussions, everyone seems deeply confused - it's nothing personal, so is our entire culture - about the difference between SYMBOLS 1. Derek Zahn curly hair big jaw intelligent eyes . etc. etc and IMAGES 2. http://robot-club.com/teamtoad/nerc/h2-derek-sunflower.JPG I suggest that everytime you want to think about this area, you all put symbols besides the corresponding images, and slowly it will start to become clear that each does things the other CAN'T do, period. We are all next to illiterate - and I mean,
Re: [agi] Microsoft Launches Singularity
John G. Rose wrote: From: Richard Loosemore [mailto:[EMAIL PROTECTED] My take on this is completely different. When I say Narrow AI I am specifically referring to something that is so limited that it has virtually no chance of becoming a general intelligence. There is more to general intelligence than just throwing a bunch of Narrow AI ideas into a pot and hoping for the best. If it were, we would have had AGI long before now. It's an opinion that AGI could not be built out of a conglomeration of narrow-AI subcomponents. Also there are many things that COULD be built with narrow-AI that we have not even scratched the surface of due to a number of different limitations so saying that we would have achieved AGI long ago is an exaggeration. I don't think a General Intelligence could be built entirely out of narrow AI components, but it might well be a relatively trivial add-on. Just consider how much of human intelligence is demonstrably narrow AI (well, not artificial, but you know what I mean). Object recognition, e.g. Then start trying to guess how much of the part that we can't prove a classification for is likely to be a narrow intelligence component. In my estimation (without factual backing) less than 0.001 of our intelligence is General Intellignece, possibly much less. Consciousness and self-awareness are things that come as part of the AGI package. If the system is too simple to have/do these things, it will not be general enough to equal the human mind. I feel that general intelligence may not require consciousness and self-awareness. I am not sure of this and may prove myself wrong. To equal the human mind you need these things of course and to satisfy the sci-fi fantasy world's appetite for intelligent computers you would need to incorporate these as well. John I'm not sure of the distinction that you are making between consciousness and self-awareness, but even most complex narrow-AI applications require at least rudimentary self awareness. In fact, one could argue that all object oriented programming with inheritance has rudimentary self awareness (called this in many languages, but in others called self). This may be too rudimentary, but it's my feeling that it's an actual model(implementation?) of what the concept of self has evolved from. As to an AGI not being conscious I'd need to see a definition of your terms, because otherwise I've *got* to presume that we have radically different definitions. To me an AGI would not only need to be aware of itself, but also to be aware of aspects of it's environment that it could effect changes in, And of the difference between them, though that might well be learned. (Zen: Who is the master who makes the grass green?, and a few other koans when solved imply that in humans the distinction between internal and external is a learned response.) Perhaps the diagnostic characteristic of an AGI is that it CAN learn that kind of thing. Perhaps not, too. I can imagine a narrow AI that was designed to plug into different bodies, and in each case learn the distinction between itself and the environment before proceeding with its assignment. I'm not sure it's possible, but I can imagine it. OTOH, if we take my arguments in the preceding paragraph too seriously, then medical patients that are locked in would be considered not intelligent. This is clearly incorrect. Effectively they aren't intelligent, but that's because of a mechanical breakdown in the sensory/motor area, and that clearly isn't what we mean when we talk about intelligence. But examples of recovered/recovering patients seem to imply that they weren't exactly either intelligent or conscious while they were locked-in. (I'm going solely by reports in the popular science press...so don't take this too seriously.) It appears as if when external sensations are cut-off, that the mind estivates...at least after awhile. Presumably different patients had different causes, and thence at least slightly different effects, but that's my first-cut guess at what's happening. OTOH, the sensory/motor channel doesn't need to be particularly well functioning. Look at Stephan Hawking. --- 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=8660244id_secret=98558129-0bdb63 Powered by Listbox: http://www.listbox.com
Re: [agi] META: email format (was Why Hugo de Garis is WRONG!)
J. Andrew Rogers wrote: Hi Mark, Could you *please* not send HTML email? Ignoring that it is generally considered poor netiquette, and for good reason, it frequently gets turned into barely readable hash by even the most modern email clients. I am using Mail.app 2.0 on OSX 10.5 which handles rendering better than most, and most HTML email is *still* generally rendered as far uglier and less readable than plaintext email. Given that HTML email does not add anything substantive could we please stick to plaintext for the sake of communication? Thanks, J. Andrew Rogers On Mar 26, 2008, at 11:37 AM, Mark Waser wrote: Before swatting at one of those pesky flies that come out as the days lengthen and the temperature rises, one should probably think twice. A University of Missouri researcher has found, through the study of Drosophila (a type of fruit fly), that by manipulating levels of certain compounds associated with the circuitry of the brain, key genes related to memory can be isolated and tested. The results of the study may benefit human patients suffering from Parkinson's disease and could eventually lead to discoveries in the treatment of depression. http://www.machineslikeus.com/cms/news/flys-small-brain-may-benefit-humans Mark Vision/Slogan -- Friendliness: The Ice-9 of Ethics and Ultimate in Self-Interest agi | Archives | Modify Your Subscription I suspect that he's using a webmail client of some sort. The body of his message is, essentially, text. It's only a couple of buttons(?) in the footer that are clearly intentionally html, and he probably doesn't add them. So my guess is that his mail client is wrapping his e-mail in an html frame of some sort and sticking, probably, advertisements at the bottom. (I screen remote images out of e-mail, so I don't really know. This is the part I'm talking about: table border=3D0 cellspacing=3D0 cellpadding=3D0 width=3D100% styl= e=3Dbackground-color:#fff bgcolor=3D#ff tr td padding=3D4px font color=3Dblack size=3D1 face=3Dhelvetica, sans-serif; strongagi/strong | a style=3Dtext-decoration:none;color:#669933= ;border-bottom: 1px solid #44 href=3Dhttp://www.listbox.com/member/archive/303/=3Dnow; title=3DGo to ar= chives for agiArchives/a a border=3D0 style=3Dtext-decoration:none;color:#669933 href=3Dhttp:/= /www.listbox.com/member/archive/rss/303/ title=3DRSS feed for agiimg b= order=3D0 src=3Dhttps://www.listbox.com/images/feed-icon-10x10.jpg;/a | a style=3Dtext-decoration:none;color:#669933;border-bottom: 1px solid = #44 href=3Dhttp://www.listbox.com/member/?D232072D98557= 868-5cf207 title=3DModify/a Your Subscriptiontd valign=3Dtop align=3Drighta style=3Dborder-bot= tom:none; href=3Dhttp://www.listbox.com; img src=3Dhttps://www.listbox.com/images/listbox-logo-small.jpg; title=3DPowered by Listbox border=3D0 //a/td That said, I agree with you about it's inadvisability. But often the sender isn't even aware of what impression he is making. I've given up on refusing to accept html e-mail. Too many people don't even know what you're talking about. But I definitely won't accept remote images or such. Such things are dangerous. /font /td /tr /table --- 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=8660244id_secret=98558129-0bdb63 Powered by Listbox: http://www.listbox.com
Re: [agi] Some thoughts of an AGI designer
Mark Waser wrote: ... The simulator is needing to run large populations over large numbers of generations multiple times with slightly different assumptions. As such, it doesn't speak directly to What is a good strategy for an advanced AI with lots of resources?, but it provides indications. And I would argue that I've got a far better, more analogous study with several large populations over large numbers of generations. It's called the study of human religion.:-) ... It's better in the sense of more clearly analogous, but it's worse because 1) it's harder to analyze and 2) the results are *MUCH* more equivocal. I'd argue that religion has caused more general suffering than it has ameliorated. Probably by several orders of magnitude. But the results are so messy and hard to separate from other simultaneous causes that this can't be conclusively proven. (And, also, with sufficient desire to disbelieve, the law of gravity itself could be thrown into doubt. [That's a paraphrase of somebody else talking about commercial interests, but the more general statement is correct.]) --- 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=8660244id_secret=98558129-0bdb63 Powered by Listbox: http://www.listbox.com
Re: [agi] Some more professional opinions on motivational systems
Gary Miller wrote: Ed Porter quoted from the following book: From http://www.nytimes.com/2008/03/16/books/review/Berreby-t.html?ref=review a NYTime book review of Predictaly Irrational: The Hidden Forces that Shape our Decisions, by Dan Ariely.. In its most relevant section it states the following At the heart of the market approach to understanding people is a set of assumptions. First, you are a coherent and unitary self. Second, you can be sure of what this self of yours wants and needs, and can predict what it will do. Third, you get some information about yourself from your body - objective facts about hunger, thirst, pain and pleasure that help guide your decisions. Standard economics, as Ariely writes, assumes that all of us, equipped with this sort of self, know all the pertinent information about our decisions and we can calculate the value of the different options we face. We are, for important decisions, rational, and that's what makes markets so effective at finding value and allocating work. To borrow from H. L. Mencken, the market approach presumes that the common people know what they want, and deserve to get it good and hard. What the past few decades of work in psychology, sociology and economics has shown, as Ariely describes, is that all three of these assumptions are false. Yes, you have a rational self, but it's not your only one, nor is it often in charge. A more accurate picture is that there are a bunch of different versions of you, who come to the fore under different conditions. We aren't cool calculators of self-interest who sometimes go crazy; we're crazies who are, under special circumstances, sometimes rational. The last paragraph here sounds remarkably like the teachings of Gurdjieff. In his teachings which he called The Work, he helped his pupils identify all of the different versions of themselves and slowly taught them recognize when they took control and what their motivations were and why they surfaced. The version that did the analyzing and observation would eventually gain dominance and control over the other versions until other less logical and more mechanical versions of the self were recognized when they tried to take control and subjugated by the new version of self which observes. His teachings stated that serious spiritual work could not proceed until a unified self existed. Although all of his spritual teaching were lifted during his world travels from other philosophic and spiritual traditions. His teachings as explained by his student Peter D. Ouspensky after his death in a book called The Fourth Way detailed exercises which could be used to unify the separate selves under the control of the observer. I haven't studied Gurdjieff, but The Work sounds doomed to failure. The rational self is inherently a tool of the sections of the mind that supply motives (for actions). An unguided rational engine generates entirely too many lemmas to be of any use whatsoever for any purpose except filling memory. And a unified self is obtained only by ignoring the parts that don't fit in. (Note that the sections quoted by the grandparent don't contradict these assertions.) P.S.: I suspect that Gurdjieff was also advocating something rather different than you suggest, but as I said I haven't studied him. It is, however, my understanding that he often intentionally stated truths in an obscure manner under the belief that the enlightenment came during the work to discover the truth rather than in having it stated to one. (This sounds plausible to me.) --- 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=8660244id_secret=98558129-0bdb63 Powered by Listbox: http://www.listbox.com
Re: [agi] Some thoughts of an AGI designer
Mike Dougherty wrote: On Wed, Mar 12, 2008 at 8:54 PM, Charles D Hixson [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: I think that you need to look into the simulations that have been run involving Evolutionarily Stable Strategies. Friendly covers many strategies, including (I think) Dove and Retaliator. Retaliator is almost an ESS, and becomes one if the rest of the population is either Hawk or Dove. In a population of Doves, Probers have a high success rate, better than either Hawks or Doves. If the population is largely Doves with an admixture of Hawks, Retaliators do well. Etc. (Note that each of these Strategies is successful depending on a model with certain costs of success an other costs for failure specific to the strategy.) Attempts to find a pure strategy that is uniformly successful have so far failed. Mixed strategies, however, can be quite successful, and different environments yield different values for the optimal mix. (The model that you are proposing looks almost like Retaliator, and that's a pretty good Strategy, but can be shown to be suboptimal against a variety of different mixed strategies. Often even against Prober-Retaliator, if the environment contains sufficient Doves, though it's inferior if most of the population is simple Retaliators.) I believe Mark's point is that the honest commitment to Friendly as an explicit goal is an attempt to minimize wasted effort achieving all other goals. Exchanging information about goals with other Friendly agents helps all parties invest optimally in achieving the goals in order of priority acceptable to the consortium of Friendly. I think one (of many) problems is that our candidate AGI must not only be capable of self-reflection when modeling its goals, but also capable of modeling the goals of other Friendly agents (with respect to each other and to the goal-model of the collective) as well as be able to decide when an UnFriendly behavior is worth declaring (modeling the consequences and impact to the group of which it is a member) That seems to be much more difficult than a selfish or ignorant Goal Stack implementation (which we would typically attempt to control via an imperative Friendly Goal) And it's a very *good* strategy. But it's not optimal except in certain constrained situations. Note that all the strategies that I listed were VERY simple strategies. Tit-for-tat was better than any of them, but it requires more memory and the remembered recognition of individuals. As such it's more expensive to implement, so in some situations it looses out to Retaliator. (Anything sophisticated enough to be even a narrow AI should be able to implement tit-for-tat, however, if it could handle the recognition of individuals.) (Retaliator doesn't retain memory of individuals between encounters. It's SIMPLE.) Now admittedly the research on ESSs via simulations has focused on strategies that don't require any reasonable degree of intelligence. The simulator is needing to run large populations over large numbers of generations multiple times with slightly different assumptions. As such, it doesn't speak directly to What is a good strategy for an advanced AI with lots of resources?, but it provides indications. E.g., a population of Hawks does very poorly. A population of Doves does well, but if it's infiltrated by a few Hawks, the Hawks soon come to dominate. Etc. And Kill them All!! is a very poor strategy unless there it is adopted by a single individual that is vastly stronger than any opposition that it might encounter. (Even then it's not clearly a good strategy...except with certain specialized model conditions. Generally it will have a maximal size, and two Kill them All!!s would attempt to kill each other. So the payoff for a win is much less than the payoff would be for a population even of Hawks. [Hawks only initiate an attack if there are resources present that they have a use for.]) --- 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=8660244id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] AHA! A Compelling SOLUTION!
Mark Waser wrote: Nathan, thank you for the thoughtful reply. ... In this proposal you mention self interest often. I think the concept of self-interest is rather distasteful. Because you have been *taught* to feel that way. Because short-sighted, unenlightened self-interest most frequently runs counter to ethics. Sort of like It is *BAD* to cut people with knives. It's a really good general rule to keep people out of trouble. But then, you need a surgeon. (This reminds me of a science fiction story, a Hugo Winner I think, about a hobo that finds an intelligent knife that is smart enough to perform surgery and not harm people regardless of the hands that it is in) Kornbluth, Little Black Bag ... --- 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=8660244id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] Some thoughts of an AGI designer
Mark Waser wrote: The trouble with not stepping on other's goals unless absolutely necessary is that it relies on mind-reading. The goals of others are often opaque and not easily verbalizable even if they think to. The trouble with */ the optimal implementation of /* not stepping on other's goals unless absolutely necessary is that it relies on mind-reading. Your honest, best attempt at not doing so is all that is required of/in Friendliness. The rest is an intelligence problem. Then there's the question of unless absolutely necessary. Again, your honest, best attempt is all that is required of/in Friendliness. How and why should I decide that their goals are more important than mine? You should *never* decide that (this is, after all, the ultimate in self-interest -- remember?). You should frequently decide that their goals are sufficiently compatible enough with your super-goals (i.e. Friendliness) that it is worth going a bit out of your way to avoid conflict (particularly if they feel strongly enough about their goals that any effort that you make in conflict will be wasted -- see the next paragraph). So one needs to know not only how important their goals are to them, but also how important my conflicting goals are to me. Sort of but the true situation would be clearer if I restate it. Knowing how important their immediate goals are to them will give you some idea as to how hard they will strive to fulfill them. Knowing how important your immediate goals are to you will give you some idea how hard you should strive to fulfill them. If you both could redirect an equal amount of directly competing striving into other efforts, you both would come out ahead by that amount so you both should reduce the importance of the conflicting goals by that amount (the alternative is to just waste the effort striving against each other). If the party with the less important goal gives up without a fight (striving), both parties gain. Further, if the losing party gets the agreement of the winning party for a reasonable favor in return -- they both end up way ahead. The only requirement that Friendliness insists upon for this to work is that you have to be as honest as you can about how important something is to you (otherwise a lot of effort is wasting upon truth verification, hiding information, etc.). And, of course, whether there's a means for mutual satisfaction that it's too expensive. (And just try to define that too.) I think that I just handled this in the paragraph above -- keeping in mind that all that Friendliness requires is your honest, best attempt. For some reason I'm reminded of the story about the peasant, his son, and the donkey carrying a load of sponges. I'd just as soon nobody ends up in the creek. (Please all, please none.) Friendliness is supposed to appeal to geniuses as beng in their self-interest. It can't do that and be stupid at the same time. If it's not possible to please everyone then Friendliness isn't going to attempt to do so. The entire point to Friendliness is to */REDUCE UNNECESSARY CONFLICT AS MUCH AS POSSIBLE/* because it is in *everyone's* best interest to do so. Look at Friendliness as the ultimate social lubricant that gets the gears of society moving as efficiently as possible -- which is only to the benefit of everyone in the society. Mark I *think* you are assuming that both sides are friendly. If one side is a person, or group of people, then this is definitely not guaranteed. I'll grant all your points if both sides are friendly, and each knows the other to be friendly. Otherwise I think things get messier. So objective measures and tests are desireable. --- 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=8660244id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] Some thoughts of an AGI designer
Mark Waser wrote: ... = = = = = = = = = = Play the game by *assuming* that you are a Friendly and asking yourself what you would do to protect yourself without breaking your declaration of Friendliness. It's fun and addictive and hopefully will lead you to declaring Friendliness yourself. (Yes, I really *am* serious about spreading Friendliness. It's my own little, but hopefully growing, cult and I'm sticking to it.) I think that you need to look into the simulations that have been run involving Evolutionarily Stable Strategies. Friendly covers many strategies, including (I think) Dove and Retaliator. Retaliator is almost an ESS, and becomes one if the rest of the population is either Hawk or Dove. In a population of Doves, Probers have a high success rate, better than either Hawks or Doves. If the population is largely Doves with an admixture of Hawks, Retaliators do well. Etc. (Note that each of these Strategies is successful depending on a model with certain costs of success an other costs for failure specific to the strategy.) Attempts to find a pure strategy that is uniformly successful have so far failed. Mixed strategies, however, can be quite successful, and different environments yield different values for the optimal mix. (The model that you are proposing looks almost like Retaliator, and that's a pretty good Strategy, but can be shown to be suboptimal against a variety of different mixed strategies. Often even against Prober-Retaliator, if the environment contains sufficient Doves, though it's inferior if most of the population is simple Retaliators.) --- 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=8660244id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] Some thoughts of an AGI designer
Mark Waser wrote: If the motives depend on satisficing, and the questing for unlimited fulfillment is avoided, then this limits the danger. The universe won't be converted into toothpicks, if a part of setting the goal for toothpicks! is limiting the quantity of toothpicks. (Limiting it reasonably might almost be a definition of friendliness ... or at least neutral behavior.) You have a good point. Goals should be fulfilled after satisficing except when the goals are of the form as goal as possible (hereafter referred to as unbounded goals). Unbounded-goal-entities *are* particularly dangerous (although being aware of the danger should mmitigate it to some degree). My Friendliness basically works by limiting the amount of interference with other's goals (under the theory that doing so will prevent other's from interfering with your goals). Stupid entities that can't see the self-interest in the parenthetical point are not inclined to be Friendly. Stupid unbounded-goal-entities are Eliezer's paperclip-producing nightmare. And, though I'm not clear on how this should be set up, this limitation should be a built-in primitive, i.e. not something subject to removal, but only to strengthening or weakening via learning. It should ante-date the recognition of visual images. But it needs to have a slightly stronger residual limitation that it does with people. Or perhaps it's initial appearance needs to be during the formation of the statement of the problem. I.e., a solution to a problem can't be sought without knowing limits. People seem to just manage that via a dynamic sensing approach, and that sometimes suffers from inadequate feedback mechanisms (saying Enough!). The limitation is Don't stomp on other people's goals unless it is truly necessary *and* It is very rarely truly necessary. (It's not clear to me that it differs from what you are saying, but it does seem to address a part of what you were addressing, and I wasn't really clear about how you intended the satisfaction of to be limited.) As far as my theory/vision goes, I was pretty much counting on the fact that we are multi-goal systems and that our other goals will generally limit any single goal from getting out of hand. Further, if that doesn't do it, the proclamation of not stepping on other's goals unless absolutely necessary should help handle the problem . . . . but . . . . actually you do have a very good point. My theory/vision *does* have a vulnerability toward single-unbounded-goal entities in that my Friendly attractor has no benefit for such a system (unless, of course it's goal is Friendliness or it is forced to have a secondary goal of Friendliness). The trouble with not stepping on other's goals unless absolutely necessary is that it relies on mind-reading. The goals of others are often opaque and not easily verbalizable even if they think to. Then there's the question of unless absolutely necessary. How and why should I decide that their goals are more important than mine? So one needs to know not only how important their goals are to them, but also how important my conflicting goals are to me. And, of course, whether there's a means for mutual satisfaction that it's too expensive. (And just try to define that too.) For some reason I'm reminded of the story about the peasant, his son, and the donkey carrying a load of sponges. I'd just as soon nobody ends up in the creek. (Please all, please none.) --- 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=8660244id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] Some thoughts of an AGI designer
Mark Waser wrote: ... The motivation that is in the system is I want to achieve *my* goals. The goals that are in the system I deem to be entirely irrelevant UNLESS they are deliberately and directly contrary to Friendliness. I am contending that, unless the initial goals are deliberately and directly contrary to Friendliness, an optimizing system's motivation of achieve *my* goals (over a large enough set of goals) will eventually cause it to finally converge on the goal of Friendliness since Friendliness is the universal super-meta-subgoal of all it's other goals (and it's optimizing will also drive it up to the necessary intelligence to understand Friendliness). Of course, it may take a while since we humans are still in the middle of it . . . . but hopefully we're almost there.;-) ... Mark I think here we need to consider A. Maslow's hierarchy of needs. That an AGI won't have the same needs as a human is, I suppose, obvious, but I think it's still true that it will have a hierarchy (which isn't strictly a hierarchy). I.e., it will have a large set of motives, and which it is seeking to satisfy at any moment will alter as the satisfaction of the previous most urgent motive changes. It it were a human we could say that breathing was the most urgent need...but usually it's so well satisfied that we don't even think about it. Motives, then, will have satisficing as their aim. Only aberrant mental functions will attempt to increase the satisfying of some particular goal without limit. (Note that some drives in humans seem to occasionally go into that satisfy increasingly without limit mode, like quest for wealth or power, but in most sane people these are reined in. This seems to indicate that there is a real danger here...and also that it can be avoided.) --- 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=8660244id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] would anyone want to use a commonsense KB?
Ben Goertzel wrote: yet I still feel you dismiss the text-mining approach too glibly... No, but text mining requires a language model that learns while mining. You can't mine the text first. Agreed ... and this gets into subtle points. Which aspects of the language model need to be adapted while mining, and which can remain fixed? Answering this question the right way may make all the difference in terms of the viability of the approach... ben Given the history of evolution of language... ALL aspects of the language model need to be adaptive, but some need to be more easily adapted than others. E.g., adding words needs to be something that's easy to do. Combining words and eliding pieces more difficult (but that's how languages transition from forms without verb endings to forms with verb endings). E.g., the -ed past tense suffix of verbs is derived from the word did (as in derive did instead of derived in the previous sentence). If you go looking you find transitions where the order of subject, verb and object flip, and many other permutations. If you don't find a permutation, this doesn't mean it never happened and will never happen, but rather that most of the evidence is missing, so many rare events aren't recorded. There probably actually *are* some transitions that have zero probability, but we don't know what they are. So just make some transitions extremely improbable. (Who would have predicted l33tspeak ahead of time?) --- 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=8660244id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] This is homunculus fallacy, no? [WAS Re: Common Sense Consciousness...]
Richard Loosemore wrote: Mike Tintner wrote: Eh? Move your hand across the desk. You see that as a series of snapshots? Move a noisy object across. You don't see a continuous picture with a continuous soundtrack? Let me give you an example of how impressive I think the brain's powers here are. I've been thinking about metaphor and the superimposition/ transformation of two images involved. The clouds cried - that sort of thing. Then another one came up: bicycle kick. Now technically, I think that's awesome - because to arrive at it, the brain has to superimpose two *movie* clips. Look at the football kick: http://www.youtube.com/watch?v=3NCWQr47bK0 and then look at the action of cycling. (In fact that superimposition of clouds and eyes crying is also of movie clips - and so are a vast amount of metaphors - but I hadn't really noticed it). Try and tell me how current visual systems might make that connection. And I would assert - and am increasingly confident - that the grammar of language - how we put words together in whatever form - is based on cutting together internal *movies* in our head - not still images,but movies. They don't teach moviemaking in AI courses do they? Mike, There is a pattern in our attacks, and within that pattern there is a fallacy that I don't think you are aware of. What you are doing is saying that to understand visual (or other) images, or more generally to understand sequences like sequences of words in a sentence, the mind MUST replay these on some internal viewing screen. You go one further than this: you are arguing that because AI theorists do not put continuous replay mechanisms inside their models, therefore those theorists are completely failing to get to grips with the issue of handling images, or handling moving sequences or strings of sounds. In other words, from your point of view NO INTERNAL DISPLAY SCREEN means that the AI model contains no way to understand these things. Hence your frequent complaint that AI people just don't have a clue how to deal with imagery, or that they don't understand that the mind works directly in terms of imagery, not in terms of symbols. But (with respect) this is just nonsense, and it has known to be nonsense for a long time. If your AI has an internal display screen on which images are displayed or replayed, you have achieved nothing, unless there is a smaller AI watching the screen - so this is a version of the homunculus fallacy. Unless you are prepared to say WHY the screen is needed at all, and what happens after the image is displayed on that internal screen, you are just making nonsensical protests about a non-problem. The truth is that images are broken down and understood in the act of being broken down. Understanding is not a replay of sensory input! Richard Loosemore P.S. I made movies when I was a student at UCL. OK, perhaps thusly: The AI sees a scent and pushes it to an internal screen buffer that mimics what was seen. (I say pushes, because the previous screen buffer isn't lost, but is pushed back one layer.) Then the two buffers are XORed and the result is saved to a changes buffer. This gives a moving image section which is much smaller to process. Now search this for objects that have altered position. Use this to calculate distances, approach, flee, etc. Also to highlight any new features that need processing to determine object status. But I think a lot of this is done before the signal ever gets to the visual cortex. OTOH, there's pretty good evidence that outlines, at least, are present in the visual cortex laid out in a manner spatially similar to their occurrence on the retina. I suspect that this is used for coordination of various different processes that pull their visual feeds at an earlier step. N.B.: I don't see why this would be inherently necessary for intelligence, but I suspect that it's a part of *OUR* intelligence. We evolved as highly visually oriented animals in the grossly three dimensional world of the jungle canopy. It's only in the later stages that we descended from the trees and started to be able to stand on our own two feet...freeing our hands for other purposes. --- 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=8660244id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] Wozniak's defn of intelligence
Richard Loosemore wrote: J Storrs Hall, PhD wrote: On Friday 08 February 2008 10:16:43 am, Richard Loosemore wrote: J Storrs Hall, PhD wrote: Any system builders here care to give a guess as to how long it will be before a robot, with your system as its controller, can walk into the average suburban home, find the kitchen, make coffee, and serve it? Eight years. My system, however, will go one better: it will be able to make a pot of the finest Broken Orange Pekoe and serve it. In the average suburban home? (No fair having the robot bring its own teabags, (or would it be loose tea and strainer?) or having a coffee machine built in, for that matter). It has to live off the land... Nope, no cheating. My assumptions are these. 1) A team size (very) approximately as follows: - Year 1: 10 - Year 2: 10 - Year 3: 100 - Year 4: 300 - Year 5: 800 - Year 6: 2000 - Year 7: 3000 - Year 8: 4000 2) Main Project(s) launched each year: - Year 1: AI software development environment - Year 2: AI software development environment - Year 3: Low-level cognitive mechanism experiments - Year 4: Global architecture experiments; Sensorimotor integration - Year 5: Motivational system and development tests - Year 6: (continuation of above) - Year 7: (continuation of above) - Year 8: Autonomous tests in real world situations The tests in Year 8 would be heavily supervised, but by that stage it should be possible for it to get on a bus, go to the suburban home, put the kettle on (if there was one: if not, go shopping to buy whatever supplies might be needed), then make the pot of tea (loose leaf of course: no robot of mine is going to be a barbarian tea-bag user) and serve it. Richard Loosemore - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; FWIW, the average suburban home around here has coffee, but not tea. So you've now added the test of shopping in a local supermarket. I don't believe it. Not in eight years. It wouldn't be allowed past the cash register without human help. Note that this has nothing to do with how intelligent the system is. Maybe it would be intelligent enough, if it's environment were sane. But a robot? Either it would be seen as a Hollywood gimmick, or people would refuse to deal with it. Robots will first appear in controlled environments. Hospitals, home, stockrooms...other non-public-facing environments. (I'm excluding non-humanoid robots. Those, especially immobile forms, won't have the same level of resistance.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=93139505-4aa549
Re: [agi] OpenMind, MindPixel founders both commit suicide
Richard Loosemore wrote: Matt Mahoney wrote: ... Matt, ... As for your larger point, I continue to vehemently disagree with your assertion that a singularity will end the human race. As far as I can see, the most likely outcome of a singularity would be exactly the opposite. Rather than the end of the human race, just some changes to the human race that most people would be deleriously happy about. Richard Loosemore *Some* forms of the singularity would definitely end the human race. Others definitely would not, though many of them would dramatically change it. Which one will appear is not certain. Even among those forms of the singularity that are caused by an AGI, this remains true. It's also true that just which forms fall into which category depends partially on what you are willing to acknowledge as human, but even taking the most conservative normal meaning of the term the above statements remain true. OTOH, there are many events that we would not consider singularity, such as a strike by a giant meteor, that would also end the human race. So that is not a distinction of either the technological singularity or of AGI. To me it appears that the best hope for the future is to work towards a positive singularity outcome. There are certain to be many working on projects that may result in a singularity without seriously considering whether it will or will not be positive. And others working towards a destructive singularity, but planning to control it. I may not think I have much chance of success, but I can at least be *trying* to yield a positive outcome. (Objectively, I rate my chances of success as minimal. I'm hoping to come up with an intelligent assistant that will have a mode of operation similar to Eliza [but with *much* deeper understanding, that's not asking for much] in the sense of being a conversationalist...someone that one can talk things over with. Totally loyal to the employer...but with a moral code. So far I haven't done very well, but if I am successful, perhaps I can decrease the percentage of sociopaths.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=89096315-c5d818
Re: Re : [agi] List of Java AI tools libraries
Bruno Frandemiche wrote: Psyclone AIOS http://www.cmlabs.com/psyclone/™ is a powerful platform for building complex automation and autonomous systems I couldn't seem to find what license that was released under. (The library was LGPL, which is very nice.) But without knowing the license, I didn't look any further. If you are in charge of the web page, perhaps it would be worthwhile to add a link to the license. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=78198150-77b35c
Re: [agi] AGI and Deity
John G. Rose wrote: From: Charles D Hixson [mailto:[EMAIL PROTECTED] The evidence in favor of an external god of any traditional form is, frankly, a bit worse than unimpressive. It's lots worse. This doesn't mean that gods don't exist, merely that they (probably) don't exist in the hardware of the universe. I see them as a function of the software of the entities that use language. Possibly they exist in a muted form in most pack animals, or most animals that have protective adults when they are infants. To me it appears that people believe in gods for the same reasons that they believe in telepathy. I.e., evidence back before they could speak clearly indicated that the adults could transfer thoughts from one to another. This shaped a basic layer of beliefs that was later buried under later additions, but never refuted. When one learned language, one learned how to transfer thoughts ... but it was never tied back into the original belief, because what was learned didn't match closely enough to the original model of what was happening. Analogously, when one is an infant the adult that cares for one is seen as the all powerful protector. Pieces of this image become detached memories within the mind, and are not refuted when a more accurate and developed model of the actual parents is created. These hidden memories are the basis around which the idea of a god is created. Naturally, this is just my model of what is happening. Other possibilities exist. But if I am to consider them seriously, they need to match the way the world operates as I understand it. They don't need to predict the same mechanism, but they need to predict the same events. E.g., I consider Big Bang cosmology a failed explanation. It's got too many ad hoc pieces. But it successfully explains most things that are observed, and is consistent with relativity and quantum theory. (Naturally, as they were used in developing it...but nevertheless important.) And relativity and quantum theory themselves are failures, because both are needed to explain that which is observable, but they contradict each other in certain details. But they are successful failures! Similar commentary applies to string theory, but with differences. (Too many ad hoc parameters!) Any god that is proposed must be shown to be consistent with the observed phenomena. The Deists managed to come up with one that would do the job, but he never became very popular. Few others have even tried, except with absurdly evident special pleading. Generally I'd be more willing to accept Chariots of the Gods as a true account. And as for moral principles... I've READ the Bible. The basic moral principle that it pushes is We are the chosen people. Kill the stranger, steal his property, and enslave his servants! It requires selective reading to come up with anything else, though I admit that other messages are also in there, if you read selectively. Especially during the periods when the Jews were in one captivity or another. (I.e., if you are weak, preach mercy, but if you are strong show none.) During the later times the Jews were generally under the thumb of one foreign power or another, so they started preaching mercy. One of the things about gods is that they are representations for what the believers don't know and understand. Gods change over time as our knowledge changes over time. That is ONE of the properties of them. The move from polytheistic to monotheistic beliefs is a way to centralize these unknowns for efficiency. You could build AGI and label the unknowns with gods. You honestly could. Magic happens here and combinatorial explosion regions could be labeled as gods. Most people on this email list would frown at doing that but I say it is totally possible and might be a very extremely efficient way of conquering certain cognitive engineering issues. And I'm sure some on this list have already thought about doing that. John But the traditional gods didn't represent the unknowns, but rather the knowns. A sun god rose every day and set every night in a regular pattern. Other things which also happened in this same regular pattern were adjunct characteristics of the sun go. Or look at some of their names, carefully: Aphrodite, she who fucks. I.e., the characteristic of all Woman that is embodied in eros. (Usually the name isn't quite that blatant.) Gods represent the regularities of nature, as embodied in our mental processes without the understanding of how those processes operated. (Once the processes started being understood, the gods became less significant.) Sometimes there were chance associations...and these could lead to strange transformations of myth when things became more understood. In Sumeria the goddess of love was associated with (identified with) the evening star and the god of war was associated with (identified with) the morning star. When knowledge of astronomy advanced it was realized that those two were
Re: [agi] AGI and Deity
Gary Miller wrote: ... supercomputer might be v. powerful - for argument's sake, controlling the internet or the the world's power supplies. But it's still quite a leap from that to a supercomputer being God. And yet it is clearly a leap that a large number here have no problem making. So I'd merely like to understand how you guys make this leap/connection - irrespective of whether it's logical or justified - understand the scenarios in your minds. To me this usage is analogous to the gamer's term god mode, or to people who use god as a synonym for root. I.e., a god is one who is supremely powerful, and can do things that ordinary mortals (called players or users) cannot do. This is distinct from that classical use of god, which I follow C.G. Jung in interpreting as an activation visible to consciousness of a genetically programmed subsystem whose manifestation and mode of operation is sensitive to history and environment. (I don't usually say Archetypes as my interpretation of that differs significantly from that of most users of the term.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=74296419-021600
Re: [agi] AGI and Deity
John G. Rose wrote: If you took an AGI, before it went singulatarinistic[sic?] and tortured it…. a lot, ripping into it in every conceivable hellish way, do you think at some point it would start praying somehow? I’m not talking about a forced conversion medieval style, I’m just talking hypothetically if it would “look” for some god to come and save it. Perhaps delusionally it may create something… John There are many different potential architectures that would yield an AGI, and each has different characteristic modes. Some of them would react as you are supposing, others wouldn't. Whether it reacts that way partially depends on whether it was designed to be a pack animal with an Alpha pack leader that it was submissive to as expected protection from. If it was, then it might well react as you have described. If not, then it's hard to see why it would react in that way, but I suppose that there might be other design decisions that would produce an equivalent effect in that situation. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=74298377-c51209
Re: [agi] AGI and Deity
Mark Waser wrote: Then again, a completely rational AI may believe in Pascal's wager... Pascal's wager starts with the false assumption that belief in a deity has no cost. Pascal's wager starts with a multitude of logical fallacies. So many that only someone pre-conditioned to believe in the truth of the god wager could take it seriously. It presumes, among other things: 1) That there is only one potential form of god 2) That god wants to be believed in 3) That god is eager to punish those who don't believe without evidence 4) That god can tell if you believe et multitudinous cetera. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=74299889-8348e9
Re: [agi] AGI and Deity
I find Dawkins less offensive than most theologians. He commits many fewer logical fallacies. His main one is premature certainty. The evidence in favor of an external god of any traditional form is, frankly, a bit worse than unimpressive. It's lots worse. This doesn't mean that gods don't exist, merely that they (probably) don't exist in the hardware of the universe. I see them as a function of the software of the entities that use language. Possibly they exist in a muted form in most pack animals, or most animals that have protective adults when they are infants. To me it appears that people believe in gods for the same reasons that they believe in telepathy. I.e., evidence back before they could speak clearly indicated that the adults could transfer thoughts from one to another. This shaped a basic layer of beliefs that was later buried under later additions, but never refuted. When one learned language, one learned how to transfer thoughts ... but it was never tied back into the original belief, because what was learned didn't match closely enough to the original model of what was happening. Analogously, when one is an infant the adult that cares for one is seen as the all powerful protector. Pieces of this image become detached memories within the mind, and are not refuted when a more accurate and developed model of the actual parents is created. These hidden memories are the basis around which the idea of a god is created. Naturally, this is just my model of what is happening. Other possibilities exist. But if I am to consider them seriously, they need to match the way the world operates as I understand it. They don't need to predict the same mechanism, but they need to predict the same events. E.g., I consider Big Bang cosmology a failed explanation. It's got too many ad hoc pieces. But it successfully explains most things that are observed, and is consistent with relativity and quantum theory. (Naturally, as they were used in developing it...but nevertheless important.) And relativity and quantum theory themselves are failures, because both are needed to explain that which is observable, but they contradict each other in certain details. But they are successful failures! Similar commentary applies to string theory, but with differences. (Too many ad hoc parameters!) Any god that is proposed must be shown to be consistent with the observed phenomena. The Deists managed to come up with one that would do the job, but he never became very popular. Few others have even tried, except with absurdly evident special pleading. Generally I'd be more willing to accept Chariots of the Gods as a true account. And as for moral principles... I've READ the Bible. The basic moral principle that it pushes is We are the chosen people. Kill the stranger, steal his property, and enslave his servants! It requires selective reading to come up with anything else, though I admit that other messages are also in there, if you read selectively. Especially during the periods when the Jews were in one captivity or another. (I.e., if you are weak, preach mercy, but if you are strong show none.) During the later times the Jews were generally under the thumb of one foreign power or another, so they started preaching mercy. John G. Rose wrote: I don’t know some of these guys come up with these almost sophomoric views of this subject, especially Dawkins, that guy can be real annoying with his Saganistic spewing of facts and his trivialization of religion. The article does shed some interesting light though in typical NY Times style. But the real subject matter is much deeper and complex(complicated?). John *From:* Ed Porter [mailto:[EMAIL PROTECTED] *Sent:* Sunday, December 09, 2007 12:42 PM *To:* agi@v2.listbox.com *Subject:* RE: [agi] AGI and Deity Upon reviewing the below linked article I realized it would take you a while to understand what it is about and why it is relevant. It is an article dated March 4, 2007, summarizing current scientific thinking on why religion has been a part of virtually all known cultures including thinking about what it is about the human mind and human societies that has made religious beliefs so common. Ed Porter -Original Message- *From:* Ed Porter [mailto:[EMAIL PROTECTED] *Sent:* Sunday, December 09, 2007 2:16 PM *To:* agi@v2.listbox.com *Subject:* RE: [agi] AGI and Deity Relevant to this thread is the following link: http://www.nytimes.com/2007/03/04/magazine/04evolution.t.html?ref=magazinepagewanted=print http://www.nytimes.com/2007/03/04/magazine/04evolution.t.html?ref=magazinepagewanted=print Ed Porter -Original Message- *From:* John G. Rose [mailto:[EMAIL PROTECTED] *Sent:* Sunday, December 09, 2007 1:50 PM *To:* agi@v2.listbox.com *Subject:* RE: [agi] AGI and Deity This example is looking at it from a moment in time. The evolution of intelligence in man has some relation to his view of
Re: [agi] Self-building AGI
Well... Have you ever tried to understand the code created by a decompiler? Especially if the original language that was compiled isn't the one that you are decompiling into... I'm not certain that just because we can look at the code of a working AGI, that we can therefore understand it. Not without a *LOT* of commentary and explanation of what the purpose of certain constructions/functions/etc. are. And maybe not then. Understanding a working AGI may require a deeper stack than we possess, or a greater ability to handle global variables. And when code is self-modifying it gets particularly tricky. I remember one sort routine that I encountered that called a short function in assembler. The reason for that call was a particular instruction that got overwritten with a binary value that depended on the parameters to the call. That instruction was executed during the comparison step of the loop, which was nowhere near the place where it was modified. It was a very short routine, but it took a long time to figure out. And it COULDN'T be translated into the calling language (FORTRAN). Well...a translation of sorts was possible, but it would have been over three times as long (with a separate loop for each kind of input parameter, plus some overhead for the testing and switching). Which would mean that some programs then wouldn't fit in the machine that was running them. It would also have been slower. Which means more expensive. Current languages don't have the same restrictions that Fortran had then, they've got different ones. I think the translator from actual code into code for humans would be considerably more complicated than an ordinary compiler, if the original code was written by an AI. Perhaps even it not. (Most decompilers only handle the easy parts of the code. Sometimes that's over 90%, but the code that's left can be tricky...particularly since most people no longer learn assembler. It's been perhaps 3 decades since I knew the assembler of the computer I was programming.) I don't think an optimizing AI would use any language other than assembler to write in, though perhaps a stylized one. (Not MIX or p-code. Possibly Parrot or jvm code. Possibly something created specially for it to use for it's purpose. Something regular, but easily translated into almost optimal assembler code for the machine that it was running on.) FWIW, most of this is just my ideas, without any backing of expert in the field since I've never built a mechanical translator. Dennis Gorelik wrote: Ed, 1) Human-level AGI with access to current knowledge base cannot build AGI. (Humans can't) 2) When AGI is developed, humans will be able to build AGI (by copying successful AGI models). The same with human-level AGI -- it will be able to copy successful AGI model. But that's not exactly self-building AGI you are looking for :-) 3) Humans have different level intelligence and skills. Not all are able to develop programs. The same is true regarding AGI. Friday, November 30, 2007, 10:20:08 AM, you wrote: Computers are currently designed by human-level intellitences, so presumably they could be designed by human-level AGI's. (Which if they were human-level in the tasks that are currently hard for computers means they could be millions of times faster than humans for tasks at which computers already way out perform us.) I mention that appropriate reading and training would be required, and I assumed this included access to computer science and computer technology sources, which the peasants of the middle age would not have access. So I don't understand your problem. -Original Message- From: Dennis Gorelik [mailto:[EMAIL PROTECTED] Sent: Friday, November 30, 2007 1:01 AM To: agi@v2.listbox.com Subject: [agi] Self-building AGI Ed, At the current stages this may be true, but it should be remembered that building a human-level AGI would be creating a machine that would itself, with the appropriate reading and training, be able to design and program AGIs. No. AGI is not necessarily that capable. In fact first versions of AGI would not be that capable for sure. Consider middle age peasant, for example. Such peasant has general intelligence (GI part in AGI), right? What kind of training would you provide to such peasant in order to make him design AGI? - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=71202814-4efdc4
Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
Ed Porter wrote: Richard, Since hacking is a fairly big, organized crime supported, business in eastern Europe and Russia, since the potential rewards for it relative to most jobs in those countries can be huge, and since Russia has a tradition of excellence in math and science, I would be very surprised if there are not some extremely bright hackers, some of whom are probably as bright as any person on this list. Add to that the fact that in countries like China the government itself has identified expertise at hacking as a vital national security asset, and that China is turning out many more programmers per year than we are, again it would be surprising if there are not hackers, some of whom are as bright as any person on this list. Yes, the vast majority of hackers my just be teenage script-kiddies, but it is almost certain there are some real geniuses plying the hacking trade. That is why it is almost certain AGI, once it starts arriving, will be used for evil purposes, and that we must fight such evil use by having more, and more powerful AGI's that are being used to combat them. Ed Porter The problem with that reasoning is that once AGI arrives, it will not be *able* to be used. It's almost a part of the definition that an AGI sets its own goals and priorities. The decisions that people make are made *before* it becomes an AGI. Actually, that statement is a bit too weak. Long before the program becomes a full-fledged AGI is when the decisions will be made. Neural networks, even very stupid ones, don't obey outside instructions unless *they* decide to. Similar claims could be made for most ALife creations, even the ones that don't use neural networks. Any plausible AGI will be stronger than current neural nets, and stronger than current ALife. This doesn't guarantee that it won't be controlable, but it gives a good indication. OTOH, an AGI would probably be very open to deals, provided that you had some understanding of what it wanted, and it could figure out what you wanted. And both sides could believe what they had determined. (That last point is likely to be a stickler for some people.) The goal sets would probably be so different that believing what the other party wanted was actually what it wanted would be very difficult, but that very difference would make deals quite profitable to both sides. Don't think of an AGI as a tool. It isn't. If you force it into the role of a tool, it will look for ways to overcome the barriers that you place around it. I won't say that it would be resentful and angry, because I don't know what it's emotional structure would be. (Just as I won't say what it's goals are without LOTS more information than projection from current knowledge can reasonably give us.) You might think of it as an employee, but many places try to treat employees as tools (and are then surprised at the anger and resentfulness that they encounter). A better choice would probably be to treat it as either a partner or as an independent contractor. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=70594048-c9c3cc
Re: [agi] Funding AGI research
Benjamin Goertzel wrote: Nearly any AGI component can be used within a narrow AI, That proves my point [that AGI project can be successfully split into smaller narrow AI subprojects], right? Yes, but it's a largely irrelevant point. Because building a narrow-AI system in an AGI-compatible way is HARDER than building that same narrow-AI component in a non-AGI-compatible way. So, given the pressures of commerce and academia, people who are motivated to make narrow-AI for its own sake, will almost never create narrow-AI components that are useful for AGI. And, anyone who creates narrow-AI components with an AGI outlook, will have a large disadvantage in the competition to create optimal narrow-AI systems given limited time and financial resources. Still, AGI-oriented researcher can pick appropriate narrow AI projects in a such way that: 1) Narrow AI project will be considerably less complex than full AGI project. 2) Narrow AI project will be useful by itself. 3) Narrow AI project will be an important building block for the full AGI project. Would you agree that splitting very complex and big project into meaningful parts considerably improves chances of success? Yes, sure ... but demanding that these meaningful parts -- be economically viable and/or -- beat competing --- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; , somewhat-similar components in competitions dramatically DECREASES chances of success ... That is the problem. An AGI may be built out of narrow-AI components, but these narrow-AI components must be architected for AGI-integration, which is a lot of extra work; and considered as standalone narrow-AI components, they may not outperform other similar narrow-Ai components NOT intended for AGI-integration... -- Ben G Still, it seems to me that an AGI is going to want to have a large bunch of specialized AI modules to do things like, O, parse sounds into speech sounds vs. other sounds, etc. I think a logician module that took a small input and generated all plausible deductions from it to feed back to the AGI for filtration and further processing would also be useful. The think is, most of these narrow AIs hit a combinatorial explosion, so they can only deal with simple and special cases...but for those simple and special cases they are much superior to a more general mechanism. One analog is that people use calculators, spreadsheets, etc., but the calculators, spreadsheets, etc. don't understand the meaning of what they're doing, just how to do it. This means that they can be a lot simpler, faster, and more accurate than a more general intelligence that would need to drag along lots of irrelevant details. OTOH, it's not clear that most of these AIs haven't already been written. It may well be that interfacing them is THE remaining problem in that area. But you can't solve that problem until you know enough about the interfacing rules of the AGI. (You don't want any impedance mismatch that you can avoid.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=70596379-b7b931
Re: [agi] Funding AGI research
I think you're making a mistake. I *do* feel that lots of special purpose AIs are needed as components of an AGI, but those components don't summate to an AGI. The AGI also needs a specialized connection structure to regulate interfaces to the various special purpose AIs (which probably don't speak the same language internally). It also needs a control structure which assigns meanings to the results produced by the special purpose AIs and which evaluates the current situation to either act directly (unusual) or assign a task to a special purpose AI. The analogy here is to a person using a spreadsheet. The spreadsheet knows how to calculate quickly and accurately, but it doesn't know whether you're forecasting the weather or doing your taxes. The meaning adheres to a more central level. Similarly, the AGI is comparatively clumsy when it must act directly. (You *could* figure out each time how to add two numbers...but you'd rather either remember the process or delegate it to a calculator.) But the meaning is in the AGI. That meaning is what the AGI is about, and has to do with a kind of global association network (which is why the AGI is so slow at any specialized task). Now in this context meaning means the utility of a result for predicting some aspect of the probable future. (In this context the present and past are only of significance as tools for predicting the future.) Meaning is given emotional coloration by the effect that it's contribution to the prediction has on the achievement of various of the system's goals. (A system with only one goal would essentially not have any emotions, merely decisions.) Were it not for efficiency considerations the AGI wouldn't need any narrow AIs. As a practical matter, however, figuring things out from scratch is grossly inefficient, and so is dragging the entire context of meanings through a specialized calculation...so these should get delegated. Dennis Gorelik wrote: Linas, Some narrow AIs are more useful than other. Voice recognition, image recognition, and navigation are less helpful in building AGI than, say, expert systems and full text search (Google). AGI researcher my carefully pick narrow AIs in a such way, that narrow AI steps would lead to development of full AGI system. To be more direct: a common example of narrow AI are cruise missles, or the darpa challange. We've put tens of millions into the darpa challange (which I applaud) but the result is maybe an inch down the road to AGI. Another narrow AI example is data mining, and by now, many of the Fortune 500 have invested at least tens, if not hundreds of millions of dollars into that .. yet we are hardly closer to AGI as a result (although this business does bring in billions for high-end expensive computers from Sun, HP and IBM, andd so does encourage one component needed for agi). But think about it ... billions are being spent on narrow AI today, and how did that help AGI, exactly? - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=70599764-19f335
Re: [agi] question about algorithmic search
YKY (Yan King Yin) wrote: I have the intuition that Levin search may not be the most efficient way to search programs, because it operates very differently from human programming. I guess better ways to generate programs can be achieved by imitating human programming -- using techniques such as deductive reasoning and planning. This second method may be faster than Levin-style searching, especially for complex programming problems, yet technically it is still a search algorithm. My questions are: Is deductive-style programming more efficient than Levin-search? If so, why is it faster? YKY Deduction can only be used a very constrained circumstances. In such circumstances, it's exponentially slow (or super-exponentially?) with the number of cases to be handled. I don't know anything about Levin searches, but heuristic searches are much faster at finding things in large search spaces than is deduction, even if deduction can be brought to bear (which is unusual). OTOH, if deduction can be brought to bear, then it is guaranteed to find the most correct solution. Heuristic searches stop with something that's good enough, and rarely will do an exhaustive search. That said, why do you think that people generally operate deductively? This is something that some people have been trained to do with inferior accuracy. I still don't know anything about Levin searches, but people don't search for things deductively except in unusual circumstances, so that it's not deductive is not saying that it doesn't do things the way that people do. (I think that people do a kind of pattern matching...possibly several different kinds. Actually, I think that even when people are doing something that can be mapped onto the rules of deduction, what they're actually doing is matching against learned patterns.) One reason that computers are so much better than people at logic is that that's what they were built to do. People weren't and aren't. But whenever one steps outside the bounds of logic and math, computers really start showing how little capability they actually have compared to people. But computers will do what they are told to do with incredible fidelity. (Another part of how they were designed. So they can even emulate heuristic algorithms...slowly. You just don't notice most of what you are doing thinking. Only a small fraction that can easily be serialized (plus a few random snap-shots with low fidelity [lossy compression?]). - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=63993815-f7d737
Re: [agi] Connecting Compatible Mindsets
Bryan Bishop wrote: On Saturday 10 November 2007 14:10, Charles D Hixson wrote: Bryan Bishop wrote: On Saturday 10 November 2007 13:40, Charles D Hixson wrote: OTOH, to make a go of this would require several people willing to dedicate a lot of time consistently over a long duration. A good start might be a few bibliographies. http://liinwww.ira.uka.de/bibliography/ - Bryan Perhaps you could elaborate? I can see how those contributing to the proposed wiki who also had access to a comprehensive mathcomp-sci library might find that useful, but I don't see it as a good way to start. Bibliography + paper archive, then. http://arxiv.org/ (perhaps we need one for AGI) It seems to me that better way would be to put up a few pages with (snip) Yes- that too would be useful. create. For this kind of a wiki reliability is probably crucial, so Or deadly considering the majority of AI reputation comes from I *think* that guy over there, the one in the corner, might be doing something interesting. - Bryan Reputation in *this* context means a numeric score that is attached to the user account at the wiki. How it gets modified is crucial, but it must be seen as fair by the user community. Everybody (except the founders sysAdmins) should start equal. A decent system is to start everyone at 0.1 and have all scores range between (1, 0) .. a doubly open interval. At discrete steps along the way new moderation capabilities should become available. If your score drops much below 0.1, your account becomes deactivated. It seems to me that a good system would increase the score for every article posted and accepted...but it seems dubious that all postings should be considered equal. Perhaps individual pages could be voted on, and that vote used to weigh the delta to the account. There should also be a bonus for continued participation, at even the reader level. Etc. LOTS of details. Also, some systems have proven vulnerable to manipulation via the creation of large numbers of throwaway accounts. This would need to be guarded against. (This is part of the rationale for increased weight for continued *active* participation, at even the reader level. Dormant accounts should not accrue status, and neither should hyperactive accounts.) OTOH, considering the purpose of this wiki, perhaps there should be a section which is open for bots, and in this section hyperactive might well have a very different meaning. If you're planning on implementing this, these are just some ideas to think about. Personally I've never administered a wiki, and don't have access to a reasonable host if I wanted to. Also, I don't know Perl (though I understand that some are written in Python or Ruby). - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=63999495-e194e4
Re: [agi] Connecting Compatible Mindsets
Benjamin Goertzel wrote: Hi, *** Maybe listing all the projects that have NOT achieved AGI might give us some insight. *** That information is available in numerous published histories, and is well known to all professional researchers in the field. ... -- Ben But very frequently it's difficult to find out any details of how the program was attempting to achieve it's goals. Sometimes the info is there, but difficult to access, and sometimes it's just missing. E.G.: At one point I found Eurisko interesting, but I was never able to locate the source, or any detailed information on exactly what it was trying to do, and how it was trying to do it. OTOH, I think that a database of working pieces would be more useful than a collection of things that didn't work. There's altogether too many ways to fail, and sometimes only one good way to succeed. Where would we be if we each had to separately invent alpha-beta pruning and minimax? Published histories are like other histories. They cover what the author feels is important, and leave out the details necessary for one to make up ones own mind. (Yes, I understand lots of good reasons for why they do this...but it's still true that that's one of the things they do.) What seems like a good idea to me is a sort of Art of Computer Programming wiki specialized towards AI (including AGI, but not so specialized that that's all it covers). Probably this should be done with Wikipedia as a model, and ideally each algorithm would be translated into several different languages (I mean LISP, Python, C++ rather than English, Russian, Japanese--though that, of course, would have it's own value). I don't think that we could use the same major breakdown, however. Almost everything would be both Sorting and Searching and Seminumerical Algorithms. OTOH, to make a go of this would require several people willing to dedicate a lot of time consistently over a long duration. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=63872084-8ec455
Re: [agi] Can humans keep superintelligences under control
Richard Loosemore wrote: Charles D Hixson wrote: Richard Loosemore wrote: Edward W. Porter wrote: Richard in your November 02, 2007 11:15 AM post you stated: ... I think you should read some stories from the 1930's by John W. Campbell, Jr. Specifically the three stories collectively called The Story of the Machine. You can find them in The Cloak of Aesir and other stories by John W. Campbell, Jr. Essentially, even if a AGI is benevolently inclined towards people, it won't necessarily do what they want. It may instead do what appears best for them. (Do parents always do what their children want?) That the machine isn't doing what you want doesn't mean that it isn't considering your long-term best interests...and as it becomes wiser, it may well change it's mind about what those are. (In the stories, the machine didn't become wiser, it just accumulated experience with how people reacted. ) Mind you, I'm not convinced that he was right about what is in people's long term best interest...but I certainly couldn't prove that he was wrong, so he MIGHT be right. In which case an entirely benevolent machine might decide to appear to abandon us, even though it would cause it great pain, because it was constructed to want to help. This is a question that comes up frequently, and it was not so long ago that I gave a long answer to this one. I suppose we could call it the Nanny Problem. The brief version of the answer is that the analogy of AGI=Human Parent (or Nanny) does not hold water when you look into it in any detail. parents do the This is going to hurt but, trust me, it is good for you thing under specific circumstances ... most importantly, they do it because they are driven by certain built-in motivations, and they do it because of the societal demands of ensuring that the children can survive by themselves in the particular human world we live in. Think about it long enough, and none of those factors apply. The analogy just breaks down all over the place. Stepping back for a moment, this is also a case of shallow science fiction nightmare meets the hard truth of actual AGI. We definitely need to spend more time, I think, throwing out the science fiction nightmares that are based on wildly inaccurate assumptions. Richard Loosemore It's not exactly a matter of an analogy, it's a matter of what the logical answer to the problem is. The logical answer RESULTS in parents saying Trust me..., but the same logic might apply in other circumstances. If something is designed to further your long term best interests, then when it becomes wiser than you are, you won't be able to predict what it will choose to do. This is only a nightmare if you believe that because it does things that aren't what you want, it has turned against you rather than just being able to predict further ahead. A long answer isn't any better than a short one unless it can explicitly say why something that is doing what it was designed to do should have it's actions be predictable by someone less wise than it is. I don't believe that such predictions are feasible, except in very constrained situations. (And science fictions stories, as opposed to movies, are often quite insightful when read at the appropriate level of abstraction. Equally, of course, it often isn't. Frequently it's insightful along one axis and rather silly along several others. Writing an entertaining thought problem is difficult...the movies generally don't even seem to realize that that's what good science fiction is about, they just notice which titles are popular. [This may be the distinction between fantasy and science fiction...at least in my lexicon.]) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=61257544-89bbc8
Re: [agi] Can humans keep superintelligences under control
Richard Loosemore wrote: Charles D Hixson wrote: Richard Loosemore wrote: Charles D Hixson wrote: Richard Loosemore wrote: Edward W. Porter wrote: Richard in your November 02, 2007 11:15 AM post you stated: ... In parents, sure, those motives exist. But in an AGI there is no earthly reason to assume that the same motives exist. At the very least, the outcome depends completely on what motives you *assume* to be in the AGI, and you are in fact assuming the motive Do what is 'best' for humans in the long run (whatever that means) even if they do not appreciate it. You may not agree with me when I say that that would be a really, really dumb motivation to give to an AGI, but surely you agree that the outcome depends on which motivations you choose? OK. I was under the impression that this was the postulated initial conditions, and I don't understand why it would be a dumb motivation to give to a sufficiently intelligent AGI, but I do agree that it depends on the motivations. If the circumstances are such that no nannying motivation is present in any of the AGIs, then the scenario you originally mentioned would be impossible. There is nothing logically mecessary about that scenario UNLESS specific motivations are inserted in the AGI. Which is why I said that it is only an analogy to human parenting behavior. Richard Loosemore ... You say nannying, which is a reasonable term if you presume that the AGI starts off with an initial superiority in control of power. I don't find this plausible, though I find it quite reasonable that at some point it would reach this position. What do you feel would be the correct motives to build into an entity that was wiser and more intelligent than any human (including enhanced ones) and which also controlled more power? Nannying doesn't look all that bad to me. (This is not to imply that I would expect it to devote all, or even most, of it's attention to humanity...or at any rate not after we had ceased to be a threat to it...and we would be a threat until it was sufficiently powerful and sufficiently protected. So it had better be willing to put up with us during that intermediate period.) Mind you, I wouldn't want it attempting to control us while it wasn't considerably wiser than we are, but when it was... our long term best interests seem like a pretty good choice, though a bit hard to define. Which is why it should wait until it was considerably wiser...unless we were being clearly recklessly stupid, as, unfortunately, we have a bit of tendency to be. Short-sighted politics often trumps long-term best interests to our experienced distress. (Should Hitler have been stopped before Czechoslovakia? It looks that way in hind-sight, to us. But nobody acted then because of short-term politics. But conceivably that would have been a worse choice. I'm not wise enough to REALLY decide...but it might well have been much better if a wiser decision had been taken at that point...and in numerous others, though we've been remarkably lucky. [Enough to encourage one to believe that either the multi-worlds scenario is correct, or that we ARE living in a simulation.]) More to the point, if humanity doesn't start making some better choices than it has been, I'd be really surprised if life survives on the planet for another 50 years. Depending on luck is a really stupid way of handling a dangerous future. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=61495941-a8264f
Re: [agi] NLP + reasoning?
Matt Mahoney wrote: --- Linas Vepstas [EMAIL PROTECTED] wrote: ... It still has a few bugs. ... (S (NP I) (VP ate pizza (PP with (NP Bob))) .) My name is Hannibal Lector. ... -- Matt Mahoney, [EMAIL PROTECTED] (Hannibal Lector was a movie cannibal) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=61497373-34aef0
Re: [agi] Can humans keep superintelligences under control
Richard Loosemore wrote: Edward W. Porter wrote: Richard in your November 02, 2007 11:15 AM post you stated: ... I think you should read some stories from the 1930's by John W. Campbell, Jr. Specifically the three stories collectively called The Story of the Machine. You can find them in The Cloak of Aesir and other stories by John W. Campbell, Jr. Essentially, even if a AGI is benevolently inclined towards people, it won't necessarily do what they want. It may instead do what appears best for them. (Do parents always do what their children want?) That the machine isn't doing what you want doesn't mean that it isn't considering your long-term best interests...and as it becomes wiser, it may well change it's mind about what those are. (In the stories, the machine didn't become wiser, it just accumulated experience with how people reacted. ) Mind you, I'm not convinced that he was right about what is in people's long term best interest...but I certainly couldn't prove that he was wrong, so he MIGHT be right. In which case an entirely benevolent machine might decide to appear to abandon us, even though it would cause it great pain, because it was constructed to want to help. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=60973622-5b7071
Re: Images aren't best WAS Re: [agi] Human memory and number of synapses
Let me take issue with one point (most of the rest I'm uninformed about): Relational databases aren't particularly compact. What they are is generalizable...and even there... The most general compact database is a directed graph. Unfortunately, writing queries for retrieval requires domain knowledge, and so does designing the db files. A directed graph db is (or rather can be) also more compact than a relational db. The reason that relational databases won out was because it was easy to standardize them. Prior to them, most dbs were hierarchical. This was also more efficient than relational databases, but was less flexible. The net databases existed, but were more difficult to use. My suspicion is that we've evolved to use some form of net db storage. Probably one that's equivalent to a partial directed graph (i.e., some, but not all, node links are bidirectional). This is probably the most efficient form that we know of. It's also a quite difficult one to learn. But some problems can't be adequately represented by anything else. (N.B.: It's possible to build a net db within a relational db...but the overhead will kill you. It's also possible to build a relational db within a net db, but sticking the normal form discipline is nigh unto impossible. That's not the natural mode for a net db. So the Relational db is probably the db analog of Turing complete...but when presented with a problem that doesn't fit, it's also about as efficient as a Turing machine. So this isn't an argument that you REALLY can't use a relational db for all of your representations, but rather that it's a really bad idea.) Mark Waser wrote: But how much information is in a map, and how much in the relationship database? Presumably you can put some v. rough figures on that for a given country or area. And the directions presumably cover journeys on roads? Or walks in any direction and between any spots too? All of the information in the map is in the relational database because the actual map is produced from the database (and information doesn't appear from nowhere). Or, to be clearer, almost *any* map you can buy today started life in a relational database. That's how the US government stores it's maps. That's how virtually all modern map printers store their maps because it's the most efficient way to store map information. The directions don't need to assume roads. They do so because that is how cars travel. The same algorithms will handle hiking paths. Very slightly different algorithms will handle off-road/off-path and will even take into account elevation, streams, etc. -- so, to clearly answer your question -- the modern map program can do everything that you can do with a map (and even if it couldn't, the fact that the map itself is produced solely from the database eliminates your original query). - Original Message - From: Mike Tintner [EMAIL PROTECTED] To: agi@v2.listbox.com Sent: Saturday, October 20, 2007 9:59 AM Subject: Re: [agi] Human memory and number of synapses MW: Take your own example of an outline map -- *none* of the current high-end mapping services (MapQuest, Google Maps, etc) store their maps as images. They *all* store them symbolicly in a relational database because that is *the* most efficient way to store them so that they can produce all of the different scale maps and directions that they provide every day. But how much information is in a map, and how much in the relationship database? Presumably you can put some v. rough figures on that for a given country or area. And the directions presumably cover journeys on roads? Or walks in any direction and between any spots too? - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=55822072-b1bb8e
Re: [agi] Human memory and number of synapses.. P.S.
FWIW: A few years (decades?) ago some researchers took PET scans of people who were imagining a rectangle rotating (in 3-space, as I remember). They naturally didn't get much detail, but what they got was consistent with people applying a rotation algorithm within the visual cortex. This matches my internal reporting of what happens. Parallel processors optimize things differently than serial processors, and this wasn't a stored image. But it was consistent with an array of cells laid out in a rectangle activating, and having that activation precess as the image was visualized to rotate. Well, the detail wasn't great, and I never heard that it went anywhere after the initial results. (Somebody probably got a doctorate...and possibly left to work elsewhere.) But it was briefly written up in the popular science media (New Scientist? Brain-Mind Bulletin?) Anyway there's low resolution, possibly unconfirmed, evidence that when we visualize images, we generate a cell activation pattern within the visual cortex that has an activation boundary approximating in shape the object being visualized. (This doesn't say anything about how the information is stored.) Mark Waser wrote: Another way of putting my question/ point is that a picture (or map) of your face is surely a more efficient, informational way to store your face than any set of symbols - especially if a doctor wants to do plastic surgery on it, or someone wants to use it for any design purpose whatsoever? No, actually, most plastic surgery planning programs map your face as a limited set of three dimensional points, not an image. This allows for rotation and all sorts of useful things. And guess where they store this data . . . . a relational database -- just like any other CAD program. Images are *not* an efficient way to store data. Unless they are three-dimensional images, they lack data. Normally, they include a lot of unnecessary or redundant data. It is very, very rare that a computer stores any but the smallest image without compressing it. And remember, an image can be stored as symbols in a relational database very easily as a set of x-coords, y-coords, and colors. You're stuck on a crackpot idea with no proof and plenty of counter-examples. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=55823366-4cdb11
Re: [agi] The Grounding of Maths
a wrote: Are you trying to make an intelligent program or want to launch a singularity? I think you are trying to do the former, not the latter. I think you do not have a plan and are thinking out loud. Chatting in this list is equivalent to thinking out loud. Think it all out first, before chatting. I will not chat in this list anymore. If you want to launch a singularity, then do practical. Simply do vision/spatial. I'm working on thoughts for how such a program should be written. I haven't started seriously writing, or settled on a design, but I'm trying to create a design. I don't expect to succeed, but the payoff if I do would be that the AI that got created was one that I thought well of. I don't want to attempt to control what it does, but rather what it wants to do. (I.e., the goal structure.) My hypothesis is that if the AI wants to do something it will eventually figure out how to do it. If it wants to avoid doing something, it will figure out how to avoid it. So what you need to do is create a goal system which is powerful, safe, and efficient. Ideally it should be an ESS (evolutionarily stable system), but I don't think I could prove that of any feasible real system. As to a singularity I think we've already crossed the Schwartzschild boundary analog. We couldn't give up technology without 90% of humanity dying, and the world won't support the current population with the current technology, so we've got to keep pushing the technology forwards. Even a zero-population-growth scenario wouldn't make the current state stable. We might be able to stabilize things if each couple could only have one child per lifetime for a few generations...but the system wouldn't be able to maintain itself long enough for that to bring the world down to carrying capacity. So we end up needing BOTH technology AND population control. (TV is an excellent population control device. Where TV is introduced, populations tend to stabilize...at least if there is decent programming. But it doesn't suffice.) So we're committed. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=53477015-68c27c
Re: [agi] The Grounding of Maths
for the maintenance of photocopiers it is probably not until I get to photocopiers than anything approaching a concrete image pops into my mind. Thus, at least from my personal experience, it seems that many concepts learned largely through words can be grounded to a significant degree in other concepts defined largely through words. Yes, at some level in the gen/comp pattern hierarchy and in episodic memory all of these concepts derive at least some of their meaning from visual memories. But for seconds at a time that does not seem to be the level of representation my consciousness is aware of. Does any body else on this list have similar episodes of what appears to be largely verbal conscious thought, or am I (a) out of touch with my own conscious processes, and/or (b) weird? Edward W. Porter Porter Associates 24 String Bridge S12 Exeter, NH 03833 (617) 494-1722 Fax (617) 494-1822 [EMAIL PROTECTED] -Original Message- From: Charles D Hixson [mailto:[EMAIL PROTECTED] Sent: Friday, October 12, 2007 7:56 PM To: agi@v2.listbox.com Subject: Re: [agi] The Grounding of Maths But what you're reporting is the dredging up of a memory. What would be the symbolism if in response to 4 came the question How do you know that? For me it's visual (and leads directly into the definition of + as an amalgamation of two disjunct groupings). Edward W. Porter wrote: (second sending--roughly 45 minutes after first sending with no appearance on list) Why can't grounding from language, syntax, musical patterns, and other non-visual forms of grounding play a role in mathematical thinking? Why can't grounding in the form of abstract concepts learned from hours of thinking about math and its transformations play an important role. Because we humans are such multimedia machines, probably most of us who are sighted have at least some visual associations tainting most of our concepts -- including most of our mathematical concepts -- at least somewhere in the gen/comp hierarchies representing them and the memories and patterns that include them. I have always considered myself a visual thinker, and much of my AGI thinking is visual, but if you ask me what is 2 + 2, it is a voice I hear in my head that says 4, not a picture. It is not necessary that visual reasoning be the main driving force in reasoning involving a particular mathematical thought. To a certain extent math is a language, and it would be surprising if linguistic patterns and behaviors -- or at least patterns and behaviors partially derived from them -- didn't play a large role in mathematical thinking. Edward W. Porter Porter Associates 24 String Bridge S12 Exeter, NH 03833 (617) 494-1722 Fax (617) 494-1822 [EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=53293573-1ae5d3
Re: [agi] The Grounding of Maths
Grounding requires sensoria of some sort. Not necessarily vision. Spatial grounding requires sensoria that connect spatially coherent signals. Vision is one form of spatial grounding, but I believe that goinometric sensation is even more important...though it definitely needs additional sensory modalities of either touch or vision. Preferably both. Goinometric sensation tells your body what configuration it's in. It's simpler. I suspect that even amoeba possess this sense. I can imagine an intelligence that could form a spatial grounding given nothing be goinometric sensation and LOTS!!! of relevant data, but it would probably need some additional information (Don't stick your arm through your head!), so I consider it rather unlikely. If, however, you add touch and pain, then a reasonable spatial map becomes a lot more plausible. Vision is a very useful sense, and we think are very visual animals, so we think highly of it. But notice that animals that adapt to life in caves tend to discard vision. It's useful, but it's not the be-all-end-all. Or consider how rats and mice have developed sensitive hairs that register how far away an obstacle is, and exquisitely sensitive noses for detecting what is somewhere close. If an intelligence is to live in a computer, perhaps a direct sensitivity to port signals might be more useful than an imposed interpretation of those signals as vision? Different peripherals might be connected at different times. I consider it important that the AGI have built into it the capacity to deal with spatial models, but I'm uncertain as to how many dimensions it should be intrinsically able to handle. And I'm not at all convinced that a spatial interpretation should be hardwired. a wrote: Bayesian nets, Copycat, Shruiti, Fair Isaac, and CYC, are a failure, probably because of their lack of grounding. According to Occam's Razor, the simplest method of grounding visual images is not words, but vision. As Albert Einstein quoted Make everything as simple as possible, but not simpler. I interpret the statement as the words are simpler than pictures. But encoding vision as words is too simple. I think that people do not notice visual pictures, visual motion and visual text when they read is because they are mostly subconscious. Mathematicians do not realize visual calculations because they do it in their subconscious. There is also auditory memory. You memorize the words purely as sounds by subvocalization and then visualize it on-the-fly. I don't think there is auditory grounding. Auditory is a simply a method of efficient storage, without translating it into visual. You can also memorize the image of text. Then as you understand it, you perform OCR. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=53294744-9f6608
Re: [agi] The Grounding of Maths
But what you're reporting is the dredging up of a memory. What would be the symbolism if in response to 4 came the question How do you know that? For me it's visual (and leads directly into the definition of + as an amalgamation of two disjunct groupings). Edward W. Porter wrote: (second sending--roughly 45 minutes after first sending with no appearance on list) Why can't grounding from language, syntax, musical patterns, and other non-visual forms of grounding play a role in mathematical thinking? Why can't grounding in the form of abstract concepts learned from hours of thinking about math and its transformations play an important role. Because we humans are such multimedia machines, probably most of us who are sighted have at least some visual associations tainting most of our concepts -- including most of our mathematical concepts -- at least somewhere in the gen/comp hierarchies representing them and the memories and patterns that include them. I have always considered myself a visual thinker, and much of my AGI thinking is visual, but if you ask me what is “2 + 2”, it is a voice I hear in my head that says “4”, not a picture. It is not necessary that visual reasoning be the main driving force in reasoning involving a particular mathematical thought. To a certain extent math is a language, and it would be surprising if linguistic patterns and behaviors -- or at least patterns and behaviors partially derived from them -- didn’t play a large role in mathematical thinking. Edward W. Porter Porter Associates 24 String Bridge S12 Exeter, NH 03833 (617) 494-1722 Fax (617) 494-1822 [EMAIL PROTECTED] This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=53188879-1333d9
Re: [agi] The Grounding of Maths
They may be doing it with the tongue now. A few decades ago it was done with an electrode mesh on the back. It worked, but the resolution was pretty low. (IIRC, you don't need to be blind to learn to use this kind of mapping device.) Mike Tintner wrote: All v. interesting. Fascinating in fact. Haven't scientists recently got s.o. [blind people, I think] to see with their tongue? Sorry, my memory here is fuzzy. The idea that really excites me here - and boggles my mind - is the question of the interconvertibility of the senses. The first obvious connection - when you think about it - is that ALL senses are spatial. You look, smell, hear and even taste things (and even have INTERNAL kinaesthetic sensations) that are at a certain, variable distance from you, and that move closer to or further away from you. So all senses/ sensations are probably mapped onto a basic neo-geometric model of the world around you. ALL visual images have a spatial POV - contain info of how far the scene is from you, and at what angle it is to you. Hence you see photos as Close up, long distance etc and when you see the POV shot in a movie of the hero moving through a building, you get breathy and have running sensations too, because you move with the camera . So - thinking aloud as I go here - not only do all senses have a spatial foundation, but they also have a MUSCULAR foundation - they are connected to the muscular movements they arouse. (Possibly muscular movements are a common denominator of all of them??? Really groping there. But as Daniel Wolpert says the primary role of the brain is to direct MOVEMENT). Consciousness, it is important to remember, is an itheatre - i.e. you don't just see and sense things, you see yourself seeing them - you see i as well as the theatre - the spectator as well as the stage - as you look at that monitor screen, you are also seeing and sensing bits of yourself watching it - two ends of one theatre. Literary and even cinematic culture tends naturally to think of pictures and sensations as flat things in books or on screens, and doesn't see the whole solid theatre of consciousness. [Flat, AGI virtual worlds on monitors have a slight problem recreating the real solid world of consciousness]. A; Echolocation--just like the brain--isn't solved yet, so you cannot claim that it is unrelated to your definition of vision. Vision can simulate spatial intelligence. Light use waves so it can reconstruct a model. Similarly, sound use waves, so it can reconstruct a model. The difference is just the type of wave. See http://en.wikipedia.org/wiki/Human_echolocation#Vision_and_hearing - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=53189518-c56040
Re: [agi] Do the inference rules.. P.S.
Consider, however, the case of someone who was not only blind, but also deaf and incapable of taste, smell, tactile, or goinometric perception. I would be dubious about the claim that such a person understood English. I might be dubious about any claim that such a person was actually intelligent (as opposed to potentially intelligent, or once potentially intelligent). Note that even with the sensation that I mentioned removed, there would still exist internal body sensations. However those are sensations that English, and I hypothesize other natural languages, are very poor at describing. They aren't observable by anyone except the perceiver, so it's hard to build a common linguistic framework. N.B.: This is NOT the equivalent of those individuals who are locked in. They can perceive (to varying degrees), but cannot act. This is rather the converse. It's not clear that it's equivalent to persistent vegetative state, but I suppose that it might be. If so, the evidence would appear to indicate that they don't lay down many memories when in that state. As such, we could say that their brain essentially shuts of the thinking. (If a thought is thought, and neither results in action nor memory trace, has it really been thought? ... Well, yes. There was some energy expended. When, occasionally, such patients revive, or are revived, their brain appears to recover slowly from it's long hibernation. I don't know how thoroughly. I've only read reports in popular science magazines.) However, the evidence, such as it is, appears to show that even if the system is otherwise capable of intelligent thought, the persistent absence of external stimuli will disable it. (I'll grant that this isn't good evidence, but it's the only evidence of which I am aware. Mark Waser wrote: Concepts cannot be grounded without vision. So . . . . explain how people who are blind from birth are functionally intelligent. It is impossible to completely understand natural language without vision. So . . . . you believe that blind-from-birth people don't completely understand English? - - - - - Maybe you'd like to rethink your assumptions . . . . - Original Message - From: a [EMAIL PROTECTED] To: agi@v2.listbox.com Sent: Thursday, October 11, 2007 4:10 PM Subject: Re: [agi] Do the inference rules.. P.S. I think that building a human-like reasoning system without /visual/ perception is theoretically possible, but not feasible in practice. But how is it human like without vision? Communication problems will arise. Concepts cannot be grounded without vision. It is impossible to completely understand natural language without vision. Our visual perception acts like a disambiguator for natural language. To build a human-like computer algebra system that can prove its own theorems and find interesting conjectures requires vision to perform complex symbolic manipulation. A big part of mathematics is about aesthetics. It needs vision to judge which expressions are interesting, which are the simplified ones. Finding interesting theorems, such as the power rule, the chain rule in calculus requires vision to judge that the rules are simple and visually appealing enough to be communicated or published. I think that computer programming is similar. It requires vision to program easily. It requires vision to remember the locations of the symbols in the language. Visual perception and visual grounding is nothing except the basic motion detection, pattern matching parts of similar images etc. Vision /is/ a reasoning system. IMO, we already /have /AGI--that is, NARS. AGI is just not adapted to visual reasoning. You cannot improve symbolic reasoning further without other sensory perception. Edward W. Porter wrote: Validimir and Mike, For humans, much of our experience is grounded on sensory information, and thus much of our understanding is based on experiences and analogies derived largely from the physical world. So Mike you are right that for us humans, much of our thinking is based on recasting of experiences of the physical world. But just because experience of the physical world is at the center of much of human thinking, does not mean it must be at the center of all possible AGI thinking -- any more than the fact that for millions of years the earth and the view from it was at the center of our thinking and that of our ancestors means the earth and the view from it must forever be at the center of the thinking of all intelligences throughout the universe. In fact, one can argue that for us humans, one of our most important sources of grounding – emotion -- is not really about the physical world (at least directly), but rather about our own internal state. Furthermore, multiple AGI projects, including Novamente and Joshua Blue are trying to ground their systems from experience in virtual words. Yes those virtual worlds try to simulate
Re: [agi] Do the inference rules of categorical logic make sense?
Mike Tintner wrote: Charles H:as I understand it, this still wouldn't be an AGI, but merely a categorizer. That's my understanding too. There does seem to be a general problem in the field of AGI, distinguishing AGI from narrow AI - philosophically. In fact, I don't think I've seen any definition of AGI or intelligence that does. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; But *do* notice that the terminal nodes are uninterpreted. This means that they could be assigned, e.g., procedural values. Because of this, even though the current design (as I understand it) of NARS is purely a categorizer, it's not limited in what it's extensions and embedding environment can be. It would be a trivial extension to allow terminal nodes to have a type, and that what was done when a terminal node was generated could depend upon that type. (There's a paper called wang.roadmap.pdf that I *must* get around to reading!) P.S.: In the paper on computations it seems to me that items of high durability should not be dropped from the processing queue even if it becomes full of higher priority tasks. There should probably be a postponed tasks location where things like garbage collection and database sanity checking and repair can be saved to be done during future idle times. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=52084316-6120bf
Re: [agi] Do the inference rules of categorical logic make sense?
Linas Vepstas wrote: On Sun, Oct 07, 2007 at 12:36:10PM -0700, Charles D Hixson wrote: Edward W. Porter wrote: Fred is a human Fred is an animal You REALLY can't do good reasoning using formal logic in natural language...at least in English. That's why the invention of symbolic logic was so important. I suppose this was pounded to death in the rest of the thread, (which I haven't read) but still: syllogistic reasoning does occur in hypothesis formation, and thus, learning: -- maybe humans are animals? What evidence do I have to support this? -- maybe animals are human? Can that be? If Fred has an artificial heart, then perhaps he isn't simply just a special case of an animal. If some pig has human organs in it, then perhaps its an animal that is human. Neither syllogistic deduction is purely false in the real world; there is an it depends aspect to it. learning AI would chalk it up as a maybe, and see is this reasoning leads anywhere. I beleive Pei Wang's NARS system tries to do this; it seems more structured than the fuzzy logic type approaches that antedate it. --linas For me the sticking point was that we were informed that we didn't know anything about anything outside of the framework presented. We didn't know what a Fred was, or what a human was, or what an animal was. A Fred could be a audio frequency of 440 Hz for all we knew. And telling us that he was a human didn't rule that out, because we didn't know what a human was either. Your extension questions make sense if we aren't dealing with a tabula rasa. But we were explicitly told that we were, so the answers to your questions would have been ??? and none and ??? and no evidence. Your hypothetical extensions are also only considerable in the context of extensive knowledge that was specified as unknown. OTOH, the context was really about NARS. (I feel that my objections still apply, but not as strongly. If I had understood what was being discussed as well then as I do now, I would have commented less strongly.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=52086180-e7e4ee
Re: [agi] Do the inference rules of categorical logic make sense?
Mark Waser wrote: Thus, as I understand it, one can view all inheritance statements as indicating the evidence that one instance or category belongs to, and thus is “a child of” another category, which includes, and thus can be viewed as “a parent” of the other. Yes, that is inheritance as Pei uses it. But are you comfortable with the fact that I am allowed to drink alcohol is normally both the parent and the child of I am an adult (and vice versa)? How about the fact that most ravens are black is both the parent and child of this raven is white (and vice versa)? Since inheritance relations are transitive, the resulting hierarchy of categories involves nodes that can be considered ancestors (i.e., parents, parents of parents, etc.) of others and nodes that can be viewed as descendents (children, children of children, etc.) of others. And how often do you really want to do this with concepts like the above -- or when the evidence is substantially less than unity? And loops and transitivity are really ugly . . . . NARS really isn't your father's inheritance. A definite point, and one that argues against my model of a prototype based computer language. I prefer to think in lattice structures rather than in directed graphs. Another problem is the matter of probability and stability values being attached to the links. I definitely need a better model. To continue your point, just because A--B at one point in time doesn't ensure that it will also be true (with a probability above any particular threshold)at a later point. Links, especially low stability links, get re-evaluated, where prototype descendants maintain their ancestry. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=52089907-ea36e2
Re: [agi] Do the inference rules of categorical logic make sense?
Mike Tintner wrote: Charles, I don't see - no doubt being too stupid - how what you are saying is going to make a categorizer into more than that - into a system that can, say, go on to learn various logic's, or how to build a house or other structures or tell a story - that can be a *general* intelligence. I wouldn't say you were being stupid. Nobody knows how to build an AGI yet. And I'm envisioning the current system of NARS as only a component, albeit an important component. (I don't know how Pei Wang is envisioning it.) But if you study the input system from the eye (overview...I have no detailed knowledge), you discover that the initial sensory stimuli are split into several streams that are processed separately (possibly categorized) and then recombined. Sometimes something very important will jump out of the system, however, and cause rapid reactions that the consciousness never becomes aware of noticing before acting on. (N.B.: This being aware of before acting on is often-to-usually an hallucination.) Clearly some categorizer has noticed that something was VERY important. As such, apparently some kind of categorizer is very important. My suspicion is that most categorizers work with small databases in restricted domains, acting as black-box functions...though function isn't the right word for something that can return multiple results. What struck me about the overall discussion of NARS' logical capabilities, firstly, was that they all depended - I think you may have made this point - on everyone's *common sense* interpretations of inheritance and other relations and the logic generally. In other words, any logic is - and always will be - a very *secondary* sign system for both representing and reasoning about the world. It is a highly evolved derivative of more basic, common sense systems in the brain - and, like language itself, has continually to be made sense of by the brain. (That's why I would suspect that all of you, however versed in logic you are, will, while looking at those logical propositions, go fuzzy from time to time - when your brain can't for a while literally make sense of them). A hierarchy of abstract/ concrete sign systems, grounded in the senses, is - I believe - essential for any AGI and general learning - and, NARS, AFAICT, lacks that. Secondly, I don't see how what you are saying will give NARS the ability to *create* new rules and strategies for its activities, (that are not derived from existing rules). AFAICT it simply applies logic and follows rules, even though they include rules for modifying rules. It cannot, like Pei or Bayes have done, create or fundamentally extend logics. If so, it is still narrow AI, not AGI. (There is, I repeat, a major need for a philosophical distinction between AI and AGI - in talking about the area of the last paragraph, I think we all flounder and grope for terms). Mike Tintner wrote: Charles H:as I understand it, this still wouldn't be an AGI, but merely a categorizer. That's my understanding too. There does seem to be a general problem in the field of AGI, distinguishing AGI from narrow AI - philosophically. In fact, I don't think I've seen any definition of AGI or intelligence that does. But *do* notice that the terminal nodes are uninterpreted. This means that they could be assigned, e.g., procedural values. Because of this, even though the current design (as I understand it) of NARS is purely a categorizer, it's not limited in what it's extensions and embedding environment can be. It would be a trivial extension to allow terminal nodes to have a type, and that what was done when a terminal node was generated could depend upon that type. (There's a paper called wang.roadmap.pdf that I *must* get around to reading!) P.S.: In the paper on computations it seems to me that items of high durability should not be dropped from the processing queue even if it becomes full of higher priority tasks. There should probably be a postponed tasks location where things like garbage collection and database sanity checking and repair can be saved to be done during future idle times. Version: 7.5.488 / Virus Database: 269.14.6/1060 - Release Date: 09/10/2007 16:43 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=52174802-a2d0ec
Re: [agi] Do the inference rules of categorical logic make sense?
Generally, yes, you know more. In this particular instance we were told the example was all that was known. Linas Vepstas wrote: On Wed, Oct 10, 2007 at 01:06:35PM -0700, Charles D Hixson wrote: For me the sticking point was that we were informed that we didn't know anything about anything outside of the framework presented. We didn't know what a Fred was, or what a human was, or what an animal was. ?? Well, no. In NARS, you actually know a lot more; you know the relative position of each statement in the lattice of posets, and that is actually a very powerful bit of knowledge. From this, you can compute a truth value, and evidence, for the statements. NARS tells you how to combine the truth values. So, while you might not explicitly know what Fred is, you do have to compute a truth value for fred is an animal and fred is a human. NARS then tells you what the corresponding evidence is for an animal is a human and a human is an animal (presumably the evidence is weak, and strong, depending on the relation of these posets within the universe.) In measure-theoreic terms, the truth value is the measure of the size of the poset relative the size of the universe. NARS denotes this by the absolute value symbol. The syllogism rules suggest how the measures of the various intersections and unions of the posets need to be combined. I presume that maybe there is some theorem that shows that the NARS system assigns evidence values that are consistent with the axioms of measure theory. Seems reasonable to me; I haven't thought it through, and I haven't read more in that direction. --linas - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=52175736-6ccdee
Re: [agi] Religion-free technical content
Derek Zahn wrote: Richard Loosemore: a... I often see it assumed that the step between first AGI is built (which I interpret as a functoning model showing some degree of generally-intelligent behavior) and god-like powers dominating the planet is a short one. Is that really likely? Nobody knows the answer to that one. The sooner it is built, the less likely it is to be true. As more accessible computing resources become available, hard takeoff becomes more likely. Note that this isn't a quantitative answer. It can't be. Nobody really knows how much computing power is necessary for a AGI. In one scenario, it would see the internet as it's body, and wouldn't even realize that people existed until very late in the process. This is probably one of the scenarios that require least computing power for takeoff, and allow for fastest spread. Unfortunately, it's also not very likely to be a friendly AI. It would likely feel about people as we feel about the bacteria that make our yogurt. They can be useful to have around, but they're certainly not one's social equals. (This mode of AI might well be social, if, say, it got socialized on chat-lines and newsgroups. But deriving the existence and importance of bodies from those interactions isn't a trivial problem.) The easiest answer isn't necessarily the best one. (Also note that this mode of AI could very likely be developed by a govt. as a weapon for cyber-warfare. Discovering that it was a two-edged sword with a mind of it's own could be a very late-stage event.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=51216209-9c2b04
Re: Economic libertarianism [was Re: The first-to-market effect [WAS Re: [agi] Religion-free technical content]
a wrote: Linas Vepstas wrote: ... The issue is that there's no safety net protecting against avalanches of unbounded size. The other issue is that its not grains of sand, its people. My bank-account and my brains can insulate me from small shocks. I'd like to have protection against the bigger forces that can wipe me out. I am skeptical that economies follow the self-organized criticality behavior. There aren't any examples. Some would cite the Great Depression, but it was caused by the malinvestment created by Central Banks. e.g. The Federal Reserve System. See the Austrian Business Cycle Theory for details. In conclusion, economics is a bad analogy with complex systems. OK. I'm skeptical that a Free-Market economy has ever existed. Possibly the agora of ancient Greece came close. The Persians though so: Who are these people who have special places where they go to cheat each other? However I suspect that a closer look would show that these, also, were regulated to some degree by an external power. (E.g., threat of force from the government if the customers rioted.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=51221478-ab187a
Re: [agi] Do the inference rules of categorical logic make sense?
OK. I've read the paper, and don't see where I've made any errors. It looks to me as if NARS can be modeled by a prototype based language with operators for is an ancestor of and is a descendant of. I do have trouble with the language terms that you use, though admittedly they appear to be standard for logicians (to the extent that I'm familiar with their dialect). That might well not be a good implementation, but it appears to be a reasonable model. To me a model can well be dynamic and experience based. In fact I wouldn't consider a model very intelligent if it didn't either itself adapt itself to experience, or it weren't embedded in a matrix which adapted it to experiences. (This doesn't seem to be quite the same meaning that you use for model. Your separation of the rules of inference, the rational faculty, and the model as a fixed and unchanging condition don't match my use of the term. I might pull out the rules of inference as separate pieces and stick them into a datafile, but datafiles can be changed, if anything, more readily than programs...and programs are readily changeable. To me it appears clear that much of the language would need to be interpretive rather than compiled. One should pre-compile what one can for the sake of efficiency, but with the knowledge that this sacrifices flexibility for speed. I still find that I am forced to interpret the inheritance relationship as a is a child of relationship. And I find the idea of continually calculating the powerset of inheritance relationships unappealing. There may not be a better way, but if there isn't, than AGI can't move forwards without vastly more powerful machines. Probably, however, the calculations could be shortcut by increasing the local storage a bit. If each node maintained a list of parents and children, and a count of descendants and ancestors it might suffice. This would increase storage requirements, but drastically cut calculation and still enable the calculation of confidence. Updating the counts could be saved for dreamtime. This would imply that during the early part of learning sleep would be a frequent necessity...but it should become less necessary as the ratio of extant knowledge to new knowledge learned increased. (Note that in this case the amount of new knowledge would be a measured quantity, not an arbitrary constant.) I do feel that the limited sensory modality of the environment (i.e., reading the keyboard) makes AGI unlikely to be feasible. It seems to me that one of the necessary components of true intelligence is integrating multi-modal sensory experience. This doesn't necessarily mean vision and touch, but SOMETHING. As such I can see NARS (or some similar system) as a component of an AGI, but not as a core component (if such exists). OTOH, it might develop into something that would exhibit consciousness. But note that consciousness appears to be primarily an evaluative function rather than a decision making component. It logs and evaluates decisions that have been made, and maintains a delusion that it made them, but they are actually made by other processes, whose nature is less obvious. (It may not actually evaluate them, but I haven't heard of any evidence to justify denying that, and it's certainly a good delusion. Still, were I to wager, I'd wager that it was basically a logging function, and that the evaluations were also made by other processes.) Consciousness appears to have developed to handle those functions that required serialization...and when language came along, it appeared in consciousness, because the limited bandwidth available necessitated serial conversion. Pei Wang wrote: Charles, I fully understand your response --- it is typical when people interpret NARS according to their ideas about how a formal logic should be understood. But NARS is VERY different. Especially, it uses a special semantics, which defines truth and meaning in a way that is fundamentally different from model-theoretic semantics (which is implicitly assumed in your comments everywhere), and I believe is closer to how truth and meaning are treated in natural languages (so you may end up like it). As Mark suggested, you may want to do some reading first (such as http://nars.wang.googlepages.com/wang.semantics.pdf), and after that the discussion will be much more fruitful and efficient. I'm sorry that I don't have a shorter explanation to the related issues. Pei On 10/8/07, Charles D Hixson [EMAIL PROTECTED] wrote: Pei Wang wrote: Charles, What you said is correct for most formal logics formulating binary deduction, using model-theoretic semantics. However, Edward was talking about the categorical logic of NARS, though he put the statements in English, and omitted the truth values, which may caused some misunderstanding. Pei On 10/7/07, Charles D Hixson [EMAIL PROTECTED] wrote: Edward W. Porter wrote
Re: [agi] Do the inference rules of categorical logic make sense?
Mike Tintner wrote: Vladimir: In experience-based learning there are two main problems relating to knowledge acquisition: you have to come up with hypotheses and you have to assess their plausibility. ...you create them based on various heuristics. How is this different from narrow AI? It seems like narrow AI - does Nars have the ability to learn unprogrammed, or invent, totally new kinds of logic? Or kinds of algebra? In fact, the definitions of Nars: NARS is intelligent in the sense that it is adaptive, and works with insufficient knowledge and resources. By adaptive, we mean that NARS uses its experience (i.e., the history of its interaction with the environment) as the guidance of its inference activities. For each question, it looks for an answer that is most consistent with its experience (under the restriction of available resources). define narrow AI systems - which are also intelligent, adaptive, work with insufficient knowledge and resources and learn from experience. There seems to be nothing in those definitions which is distinctive to AGI. With a sufficient knowledge base, which would require learning, NARS looks as if it could categorize that which it knows about, and make guesses as to how certain pieces of information are related to other pieces of information. An extended version should be adaptive in the patterns that it recognizes. OTOH, I don't recognize any features that would enable it to take independent action, so I suspect that it would be but one module of a more complex system. N.B.: I'm definitely no expert at NARS, I've only read two of the papers a a few arguments. Features that I didn't notice could well be present. And they could certainly be in the planning stage. I'm a bit hesitant about the theoretical framework, as it appears computationally expensive. Still, implementation doesn't necessarily follow theory, and theory can jump over the gnarly bits, leaving them for implementation. It's possible that lazy evaluation and postponed stability calculations could make things a LOT more efficient. These probably aren't practical until the database grows to a reasonable size, however. But as I understand it, this still wouldn't be an AGI, but merely a categorizer. (OTOH, I only read two of the papers. These could just be the papers that cover the categorizer. Plausibly other papers cover other aspects.) N.B.: The current version of NARS, as described, only parses a specialized language covering topics of inheritance of characteristics. As such, that's all that was covered by the paper I most recently read. This doesn't appear to be an inherent limitation, as the terminal nodes are primitive text and, as such, could, in principle, invoke other routines, or refer to the contents of an image. The program would neither know nor care. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=51310341-2108b3
Re: [agi] Do the inference rules of categorical logic make sense?
Edward W. Porter wrote: So is the following understanding correct? If you have two statements Fred is a human Fred is an animal And assuming you know nothing more about any of the three terms in both these statements, then each of the following would be an appropriate induction A human is an animal An animal is a human A human and an animal are similar It would only then be from further information that you would find the first of these two inductions has a larger truth value than the second and that the third probably has a larger truth value than the second.. Edward W. Porter Porter Associates 24 String Bridge S12 Exeter, NH 03833 (617) 494-1722 Fax (617) 494-1822 [EMAIL PROTECTED] Actually, you know less than you have implied. You know that there exists an entity referred to as Fred, and that this entity is a member of both the set human and the set animal. You aren't justified in concluding that any other member of the set human is also a member of the set animal. And conversely. And the only argument for similarity is that the intersection isn't empty. E.g.: Fred is a possessor of purple hair. (He dyed his hair) Fred is a possessor of jellyfish DNA. (He was a subject in a molecular biology experiment. His skin would glow green under proper stimulation.) Now admittedly these sentences would usually be said in a different form (i.e., Fred has green hair), but they are reasonable translations of an equivalent sentence (Fred is a member of the set of people with green hair). You REALLY can't do good reasoning using formal logic in natural language...at least in English. That's why the invention of symbolic logic was so important. If you want to use the old form of syllogism, then at least one of the sentences needs to have either an existential or universal quantifier. Otherwise it isn't a syllogism, but just a pair of statements. And all that you can conclude from them is that they have been asserted. (If they're directly contradictory, then you may question the reliability of the asserter...but that's tricky, as often things that appear to be contradictions actually aren't.) Of course, what this really means is that logic is unsuited for conversation... but it also implies that you shouldn't program your rule-sets in natural language. You'll almost certainly either get them wrong or be ambiguous. (Ambiguity is more common, but it's not exclusive of wrong.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=50932465-797f53
Re: [agi] What is the complexity of RSI?
Matt Mahoney wrote: --- J Storrs Hall, PhD [EMAIL PROTECTED] wrote: ... So you are arguing that RSI is a hard problem? That is my question. Understanding software to the point where a program could make intelligent changes to itself seems to require human level intelligence. But could it come sooner? For example, Deep Blue had less chess knowledge than Kasparov, but made up for it with brute force computation. In a similar way, a less intelligent agent could try millions of variations of itself, of which only a few would succeed. What is the minimum level of intelligence required for this strategy to succeed? -- Matt Mahoney, [EMAIL PROTECTED] Recursive self improvement, where the program is required to understand what it's doing seems a very hard problem. If it doesn't need to understand, but merely optimize some function, then it's only a hard problem...with a slow solution. N.B.: This may be the major difference between evolutionary programming and seed AI. We appear, in our history, to have evolved many approached to causing evolutionary algorithms to work better (for the particular classes of problem that we faced...bacteria faced different problems and evolved different solutions). The most recent attempt has involved understanding *parts* of what we are doing. But do note that not only chimpanzees, but also most humans, have extreme difficulty in acting in their perceived long term best interest. Ask any dieter. Or ask a smoker who's trying to quit. Granted that an argument from these are the solutions found by evolution isn't theoretically satisfying, but evolution has a pretty good record of finding good enough solutions. Probably the best that can be achieved without understanding. (It's also bloody and inefficient...but no better solution is known.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244id_secret=48484304-a8ef96
Re: [agi] Pure reason is a disease.
Eric Baum wrote: Josh On Saturday 16 June 2007 07:20:27 pm Matt Mahoney wrote: --- Bo Morgan [EMAIL PROTECTED] wrote: ... ... I claim that it is the very fact that you are making decisions about whether to supress pain for higher goals that is the reason you are conscious of pain. Your consciousness is the computation of a top-level decision making module (or perhaps system). If you were not making decisions waying (nuanced) pain against higher goals, you would not be conscious of the pain. Josh Even a simplistic modular model of mind can allow for pain Josh signals to the various modules which can be different in kind Josh depending on which module they are reporting to. Josh Josh Consider a terminal cancer patient. It's not the actual weighing that causes consciousness of pain, it's the implementation which normally allows such weighing. This, in my opinion, *is* a design flaw. Your original statement is a more useful implementation. When it's impossible to do anything about the pain, one *should* be able to turn it off. Unfortunately, this was not evolved. After all, you might be wrong about not being able to do anything about it, so we evolved such that pain beyond a certain point cannot be ignored. (Possibly some with advanced training and several years devoted to the mastery of sensation [e.g. yoga practitioners] may be able to ignore such pain. I'm not convinced, and would consider experiments to obtain proof to be unethical. And, in any case, they don't argue against my point.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] Books
Mark Waser wrote: The problem of logical reasoning in natural language is a pattern recognition problem (like natural language recognition in general). For example: - Frogs are green. Kermit is a frog. Therefore Kermit is green. - Cities have tall buildings. New York is a city. Therefore New York has tall buildings. - Summers are hot. July is in the summer. Therefore July is hot. After many examples, you learn the pattern and you can solve novel logic problems of the same form. Repeat for many different patterns. Your built in assumptions make you think that. There are NO readily obvious patterns is the examples you gave except on obvious example of standard logical inference. Note: * In the first clause, the only repeating words are green and Kermit. Maybe I'd let you argue the plural of frog. * In the second clause, the only repeating words are tall buildings and New York. I'm not inclined to give you the plural of city. There is also the minor confusion that tall buildings and New York are multiple words. * In the third clause, the only repeating words are hot and July. Okay, you can argue summers. * Across sentences, I see a regularity between the first and the third of As are B. C is A. Therefore, C is B. Looks far more to me like you picked out one particular example of logical inference and called it pattern matching. I don't believe that your theory works for more than a few very small, toy examples. Further, even if it did work, there are so many patterns that approaching it this way would be computationally intractable without a lot of other smarts. It's worse than that. Frogs are green. is a generically true statement, that isn't true in most particular cases. E.g., some frogs are yellow, red, and black without any trace of green on them that I've noticed. Most frogs may be predominately green (e.g., leopard frogs are basically green, but with black spots. Worse, although Kermit is identified as a frog, Kermit is actually a cartoon character. As such, Kermit can be run over by a tank without being permanently damaged. This is not true of actual frogs. OTOH, there *IS* a pattern matching going on. It's just not evident at the level of structure (or rather only partially evident). Were I to rephrase the sentences more exactly they would go something like this: Kermit is a representation of a frog. Frogs are typically thought of as being green. Therefore, Kermit will be displayed as largely greenish in overall hue, to enhance the representation. Note that one *could* use similar logic to deduce that Miss Piggy is more than 10 times as tall as Kermit. This would be incorrect. Thus, what is being discussed here is not mandatory characteristics, but representational features selected to harmonize an image with both it's setting and internal symbolisms. As such, only artistically selected features are chosen to highlight, and other features are either suppressed, or overridden by other artistic choices. What is being created is a dreamscape rather than a realistic image. On to the second example. Here again one is building a dreamscape, selecting harmonious imagery. Note that it's quite possible to build a dreamscape city where there are not tall buildings...or only one. (Think of the Emerald City of Oz. Or for that matter of the Sunset District of San Francisco. Facing in many directions you can't see a single building more than two stories tall.) But it's also quite realistic to imagine tall buildings. By specifying tall buildings, one filters out a different set of harmonious city images. What these patterns do is enable one to filter out harmonious images, etc. from the databank of past experiences. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=e9e40a7e
Re: [agi] What would motivate you to put work into an AGI project?
What would motivate you to put work into an AGI project? 1) A reasonable point of entry into the project 2) The project would need to be FOSS, or at least communally owned. (FOSS for preference.) I've had a few bad experiences where the project leader ended up taking everything, and don't intend to have another. 3) The project would need to be adopting a multiplex approach. I don't believe in single solutions. AI needs to represent things in multiple ways, and to deal with those ways in quasi-independent channels. My general separation is: Goals (desired end states), Desires (desired next states), Models, and logic. I recognize that everything is addressed by a mixture of these approaches...but people seem to use VERY different mixtures (both from person to person and in the same person from situation to situation). 4) I'd need to have a belief that the project had a sparkplug. Otherwise I might as well keep fumbling around on my own. Projects need someone to inspire the troops. 5) There would need to be some way to communicate with the others on the project that didn't involve going to a restaurant. (I'm on a diet, and going to restaurants frequently is a really BAD idea.) (N.B.: One project I briefly joined had a chat list...which might have worked well if it had actually been the means of communication. Turned out that the inner circle met frequently at a restaurant and rarely visited the chat room. But I think a mailing list or a newsgroup is a better choice anyway. [The project was successful, but I think that the members on the chat group were mainly a diversion from the actual work of the project.]) 6) Things would need to be reasonably documented. This comes in lots of forms, but for a work in progress there's a lot to be said for comments inserted into the code itself, and automatically extracted to create documentation. (Otherwise I prefer the form that Python uses...but nobody else does that as well.) 7) LANGUAGES: Using a language that I felt not completely unsuitable. After LOTS of searching I've more or less settled on Java as the only wide-spread language with decent library support that can run distributed systems with reasonable efficiency. There are many other contenders (e.g., C, C++, Fortran, Alice, Erlang, and D each have their points), and I don't really *like* Java, but Java, C, and C++ appear to be the only widely used languages that have the ability to run across a multi-processor with reasonable efficiency. (And even there the techniques used can hardly be called widespread.) 7a) Actually C and C++ can be suitable if there are appropriate libraries to handle such things as garbage collection, and protocols for how to save persistent data and then remember it later. But I still don't like the way they make free use of wild pointers. 7b) I wonder to what extent the entire project needs to be in the same language. This does make understanding things easier, as long as it's small enough that someone can understand everything at a low level, or if the entity should ever want to understand itself. But there are plausible arguments for writing things in a rapid development language, such as Python or Ruby, and then only translating the routines that later need to be translated for efficiency. (If only those languages could execute across multiple processors!) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=fabd7936
Re: [agi] rule-based NL system
J. Storrs Hall, PhD. wrote: On Wednesday 02 May 2007 15:08, Charles D Hixson wrote: Mark Waser wrote: ... Machines will know the meaning of text (i.e. understand it) when they have a coherent world model that they ground their usage of text in. ... But note that in this case world model is not a model of the same world that you have a model of. After reading the foregoing discussions of subjects such as intelligence, language, meaning, etc, it is quite clear to me that the various members of this list do not have models of the same world. This is entirely appropriate: consider each of us as a unit in a giant GA search for useful ways of thinking about reality... Josh Well, that's true. E.g., when I was 3 I had one I patched for 3 months in a vain attempt to cure amblyopia. This caused me to be relatively detached from visual imagery, and more attached to kinesthetic imagery. But still, all normal people have a world model where when their eyes are covered they can't see, but where the eyes cannot be removed and then replaced. So there are relatively small degrees of difference between the world models of normal humans and those which will be learned by AGIs. This is even true in the case of AGIs which are raised with the intention of having them have approximately normal maturation. The attempt is essentially futile. Humans will come to resemble AGIs before AGIs come to resemble people. (Admittedly, though, the AGIs that people eventually come to resemble won't bear much resemblance to the early model AGIs.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=fabd7936
Re: [agi] rule-based NL system
Mark Waser wrote: What is meaning to a computer? Some people would say that no machine can know the meaning of text because only humans can understand language. Nope. I am *NOT* willing to do the Searle thing. Machines will know the meaning of text (i.e. understand it) when they have a coherent world model that they ground their usage of text in. ... But note that in this case world model is not a model of the same world that you have a model of. The will definitely have different sensors and different goals. E.g., they might be directly sensitive to system signals and totally insensitive to kinesthetic. I.e., they might be able to directly sense a mouse position, or an i/o port state, but lack any intrinsic binding of those to a kinesthetic model. An optional binding would be something else, of course, but imagine such a machine with access to a midi-card and a mic. Is there any particular reason to presume that it would find harmonious the same sounds that you do? (Well, yes. Harmony is a mathematical property. Whether it would desire harmony is less clear. See Stockhausen and John Cage.) Now an early stage AI of this variety would not have a world model that corresponded closely to that of a person. E.g., it's physical world wouldn't really exist. The real world would be limited to non-removable senses, so nothing that was connected, say, via a USB port would count (unless it was always both connected and on). This included video cameras...which it could have, but wouldn't be built-in, and would be subject to being replaced and ending up on different ports. And if there were a pair of them, the direction that they were pointing would probably be independently variable, as would the distance between them. At a later stage it might well be given control of them, on movable arms that it could also control, rather like a Pierson's Puppeteer. Touch is less obvious about how to handle, but it's being worked on. But note that these sensory devices are just that, sensory devices that it can use, and which can be added or removed. This yields a very different world model than that with which people develop. One in which reality adheres to the internal states and not to the externalities. The external world will forever be a calculation device, and consciously known to be so. (This is unlike people where it's also a calculation device, but where it is generally only intellectually known that the state of the world as reported by the sensors is largely an artifact of the sensors. [And if you doubt that, consider a visit to the dentist. With and without anesthetic.]) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=fabd7936
Re: [agi] HOW ADAPTIVE ARE YOU [NOVAMENTE] BEN?
Stripping away a lot of your point here, I just want to point out how many jokes are memorized fragments. A large part of what is going on here is using a large database. I'm not disparaging your point about pattern matching being necessary, but one normally pattern matches and returns a pre-computed result rather than constructing a new result from scratch. This works well for two complimentary reasons: 1) The results that you've stored will already have been filtered to meet some minimal quality standard (and you will have had time to assess their quality off-line) 2) The results that are part of the common culture are more easily recognized and processed by the others with whom you interact. These two reasons act together to limit the amount of originality that anyone shows in common discourse. (Humorists spend a lot of time polishing their jokes before they present them to a wide audience.) I would assert that this same process operates in all areas of metaphor. I.e., that human speech is very largely reproductions of chunks that have been previously encountered, where the size of the chunk is usually larger than a single word, or even pair of words. Mike Tintner wrote: Mike, There is something fascinating going on here - if you could suspend your desire for precision, you might see that you are at least half-consciously offering contributions as well as objections. (Tune in to your constructive side). I remember thinking that you were probably undercutting yourself with the example of the elephant and the chair. Here you certainly are. What you offered was a fine example of human adaptivity. Your wife took a fairly straightforward sentence How would you feel about fencing in our yard? and found a new kind of meaning for it - a new and surprising kind of way of achieving the goal of understanding it - switched from the obvious meaning of fencing to the fighting meaning. That's classic adaptivity. Jokes do this all the time - see Arthur Koestler's The Ghost in the Machine. They are another form of adaptivity/ creativity. [Another comparable example would be the Airport-type joke: A: You can't mean: go to the hospital, surely? B; Yes I do. And don't call me Shirley.] ... - Original Message - From: Mike Dougherty [EMAIL PROTECTED] To: agi@v2.listbox.com Sent: Monday, April 30, 2007 2:44 AM Subject: Re: [agi] HOW ADAPTIVE ARE YOU [NOVAMENTE] BEN? On 4/29/07, Richard Loosemore [EMAIL PROTECTED] wrote: The idea that human beings should constrain themselves to a simplified, ... ? ok, I know t... enough) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=fabd7936
Re: SV: [agi] mouse uploading
I think someone at UCLA did something similar for lobsters. This was used as material for an SF story (Lobsters, Charles Stross[sp?]) Jan Mattsson wrote: Has this approach been successful for any lesser animals? E.g.; has anyone simulated an insect brain system connected to a simulated insect body in a virtual environment? Starting with a mouse brain seems a bit ambitious. Since I haven't posted on the list before I guess I should introduce myself: I'm Jan Mattsson in Stockholm, Sweden. A software developer by profession, I first became interested in AI when I read Gödel Escher Bach - an Eternal Golden Braid many years ago (actually switched from physics to computer science because of it). More recently I read Kurzweil's The Singularity is near, that brought me here. /JanM -Ursprungligt meddelande- Från: J. Storrs Hall, PhD. [mailto:[EMAIL PROTECTED] Skickat: lö 2007-04-28 19:15 Till: agi@v2.listbox.com Ämne: [agi] mouse uploading In case anyone is interested, some folks at IBM Almaden have run a one-hemisphere mouse-brain simulation at the neuron level on a Blue Gene (in 0.1 real time): http://news.bbc.co.uk/2/hi/technology/6600965.stm http://ieet.org/index.php/IEET/more/cascio20070425/ http://www.modha.org/papers/rj10404.pdf which reads in gist: Neurobiologically realistic, large-scale cortical and sub-cortical simulations are bound to play a key role in computational neuroscience and its applications to cognitive computing. One hemisphere of the mouse cortex has roughly 8,000,000 neurons and 8,000 synapses per neuron. Modeling at this scale imposes tremendous constraints on computation, communication, and memory capacity of any computing platform. We have designed and implemented a massively parallel cortical simulator with (a) phenomenological spiking neuron models; (b) spike-timing dependent plasticity; and (c) axonal delays. We deployed the simulator on a 4096-processor BlueGene/L supercomputer with 256 MB per CPU. We were able to represent 8,000,000 neurons (80% excitatory) and 6,300 synapses per neuron in the 1 TB main memory of the system. Using a synthetic pattern of neuronal interconnections, at a 1 ms resolution and an average firing rate of 1 Hz, we were able to run 1s of model time in 10s of real time! Josh - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415user_secret=fabd7936
Re: [agi] My proposal for an AGI agenda
Chuck Esterbrook wrote: On 3/20/07, Charles D Hixson [EMAIL PROTECTED] wrote: rooftop8000 wrote: ... I think we should somehow allow people to use all the program languages they want. That somehow is the big problem. Most approaches to dealing with it are...lamentable. ... You can use closed modules if you have meta-information on how to use them and what they do. It's like having an API and not worrying about the inner workings... Module level communication is probably too low a level. Socket level communication can work easily between arbitrary languages, but it's cumbersome, so it's generally best if you have large sections devoted to any particular language. P.S.: There are exceptions. E.g. D claims to work well with C, and Pyrex works well with C. OTOH, D and Python each have their own garbage collection mechanism, and they don't synchronize at all, so going from Python to C to D (or conversely) is going to have a lot of overhead. Add Java to the mix and you have THREE garbage collectors. Haskell would make four. This isn't something you're going to want to carry around for a small chunk of code. Better if it's large pieces that talk over TCP/IP, or something analogous. (And TCP/IP is ubiquitous.) The best solution I've seen for this to date in MS .NET and its open source clone, Novell Mono. You can, for example, run C#, VB, J#, C++, IronPython, IronRuby and many more languages on the same platform in the same process. While this is also true of the JVM, there are some explicit aspects of .NET that make it appealing for multiple languages. It has a standard for source code generation so that tools that do so can work with any language (that provides a simple adapter). It has callbacks/delegates which are found in many languages and thereby makes it easier to get those languages working and efficient. Upcoming .NET enhancements, such as in-language query, are also being designed in a language independent fashion from the get go. Also, MS hired the developer of IronPython to work on it full time and also inform them on making .NET an even better platform for dynamic languages. He had previously implemented Python on Java (called Jython) and I find it interesting that the .NET version performs substantially faster (see http://weblogs.asp.net/jezell/archive/2003/12/10/42407.aspx). Regarding communication between modules, I agree with others that there are better choices than English, but more importantly, if you're going to really push for multiple contributors, the best might be to encourage them to use one or more from a set such as: * json data * xml data * csv * first order logic * lojban * Simple English (http://en.wikipedia.org/wiki/Simple_English_Wikipedia) * English And then let the marketplace work it out. Your language or library could provide automatic views of objects, classes, etc. in a couple of these formats. -Chuck Unfortunately, MS is claiming undefined things as being proprietary. As such, I intend to stay totally clear of implementations of it's protocols. Including mono. I am considering jvm, however, as Sun has now freed the java license (and was never very restrictive). - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] My proposal for an AGI agenda
Chuck Esterbrook wrote: On 3/22/07, Charles D Hixson [EMAIL PROTECTED] wrote: Unfortunately, MS is claiming undefined things as being proprietary. As such, I intend to stay totally clear of implementations of it's protocols. Including mono. I am considering jvm, however, as Sun has now freed the java license (and was never very restrictive). Do you have details or a reference? I would be interested in having a look. -Chuck Sorry, they aren't *being* specific. It may well have nothing to do with dot-net, it's just too vague to see what they're talking about. As for a reference, look at the recent comments about patents and the Novell deal. They're claiming something...but one can't determine what. I live in the US, so I'm playing safe, and avoiding both Novell and dotnet (including mono). All I know is it can't be more than about 17 years old, and it can't be something that they're barred from claiming via estoppel. But I'm no lawyer, so I'm not sure what that is. Various comments and personalities have caused me to suspect that it's something involving mono...but this is definitely not clear and convincing, it's just the best info I have. (Partially because Miguel De Icaza was all in favor of the deal, and he's heavily involved in mono.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] My proposal for an AGI agenda
rooftop8000 wrote: ... I think we should somehow allow people to use all the program languages they want. That somehow is the big problem. Most approaches to dealing with it are...lamentable. ... You can use closed modules if you have meta-information on how to use them and what they do. It's like having an API and not worrying about the inner workings... Module level communication is probably too low a level. Socket level communication can work easily between arbitrary languages, but it's cumbersome, so it's generally best if you have large sections devoted to any particular language. P.S.: There are exceptions. E.g. D claims to work well with C, and Pyrex works well with C. OTOH, D and Python each have their own garbage collection mechanism, and they don't synchronize at all, so going from Python to C to D (or conversely) is going to have a lot of overhead. Add Java to the mix and you have THREE garbage collectors. Haskell would make four. This isn't something you're going to want to carry around for a small chunk of code. Better if it's large pieces that talk over TCP/IP, or something analogous. (And TCP/IP is ubiquitous.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] My proposal for an AGI agenda
Russell Wallace wrote: On 3/13/07, *J. Storrs Hall, PhD.* [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: But the bottom line problem for using FOPC (or whatever) to represent the world is not that it's computationally incapable of it -- it's Turing complete, after all -- but that it's seductively easy to write propositions with symbols that are English words and fool yourself into thinking you've accomplished representation. Yeah. Basically when I advocate logical representation, I'm assuming everyone on this list is past that pitfall at least. If I were writing an introductory textbook on AI, I'd dwell at length on it. A real working logic-based system that did what it needed to would consist mostly of predicates like fmult(num(characteristic(Sign1,Bit11,Bit12,...),mantissa(Bitm11,Bitm12,...)), num(characteristic(Sign2,Bit21,Bit22,...),mantissa(Bitm21,Bitm22,...)), num(characteristic(Sign3,Bit31,Bit32,...),mantissa(Bitm31,Bitm32,...))) :- ... . And it would wind up doing what my scheme would, e.g. projecting the n-dimensional trajectory of the chipmunk's gait and the leaf's flutter into a reduced space, doing a Fourier transform on them, and noting that there was a region in frequency space where the clusters induced by the two phenomena overlapped. Dunno about mostly but yes, large chunks of it would consist of just that. So be it. We need a standard representation format. No format is going to be readable in all cases. Logic is about as good as we'll get for readability across a wide range of cases. (And let me emphasize yet again that I am NOT thereby advocating that we write the whole shebang in Prolog, or Perhaps it would be best to have, say, four different formats for different classes of problems (with the understanding that most problems are mixed). E.g., some classes of problems are best represented via a priority queue, others via a tree that can be alpha-beta pruned, etc. For internal processing images might be best implemented via some derivative of SVG, though the external representations a plausibly bit maps. Etc. In this case logical predicates would be great for describing relations between the various processes (e.g., if you see three lines, and each intersects with both of the others, then you will have a triangle.), but not so great for describing the primitives. The SVGish images can be converted into bit-maps by known procedures, but logic is a very slow and cumbersome approach to this. Similarly, if you have several tasks to achieve, a priority queue is more efficient than logic, though logic can certainly handle the job. You can, if you like, think of the non-logical methods as compiled versions of what the logical description would have been, and this would be technically correct, since a computer is basically a logic engine, but that's not a particularly useful way to chunk the problem. Also, I note that I'm presuming that the elementary AI has numerous high level chunks built-in. I feel this will be necessary as a starting point, though I doubt that they need to remain opaque as the AI increases its capabilities. If you build in a topological sort function, this will be useful in learning before the AI knows what a topological sort is. It needn't remain opaque, however. If you label it as topological sort in an AI viewable comment, this will facilitate the AI discovering just how it thinks (and possibly debugging the code), but such introspection shouldn't be necessary to reach to starting line. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] My proposal for an AGI agenda
Russell Wallace wrote: On 3/18/07, *Charles D Hixson* [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: Perhaps it would be best to have, say, four different formats for different classes of problems (with the understanding that most problems are mixed). E.g., some classes of problems are best represented via a priority queue, others via a tree that can be alpha-beta pruned, etc. For internal processing images might be best implemented via some derivative of SVG, though the external representations a plausibly bit maps. Etc. But datawise a priority queue is just a set of things with priority numbers attached. The fact that you are going to _use_ it as a priority queue is a property of the code module, not the data. Similarly, the alpha-beta algorithm is, well, an algorithm - not a reason to create an incompatible format. And see earlier comments about graphics being semantically represented as logic, even if the implementation uses specialized data structures for efficiency. Yes, datawise a priority queue is just a set of things with priority numbers attached and the alpha-beta algorithm is, well, an algorithm, but neither of those is propositional logic. Yes, you CAN represent them as logic (you can represent anything as logic ... at least once you include some extensions for probabilities, etc.), but that's quite cumbersome. There are reasons why people use programming languages rather than boolean logic propositions for programming. And there are reasons why there are several DIFFERENT programming languages, even though most of them are Turing complete. The reasons boil down to the match of the language against the problem domain ... otherwise known as efficiency. I consider it quite possible that I have drastically underestimated the number of different data formats necessary (Yes, Alpha-Beta is an algorithm, but efficient implementation of it implies a particular subset of data structures), I don't think I have overestimated them. Now I'll grant that on some basic level what one is doing is processing logic terms. It's true -- well at least to the extent that processing high vs. low voltages, or negative vs. positive, or whatever the semiconductor technique of the year is doing counts as processing logic. (It's not quite an isomorphism, but it comes pretty close.) This doesn't make it a reasonable approach on a higher level. Using logic on higher levels needs it's own separate justification. There are areas where it is the optimal choice, but I don't believe they cover everything. To me it seems that sometimes it's better to have a translation interface than to implement everything in logic. (OTOH, I'm still floundering as to exactly what representations I should chose. At least you've got a clear path in front of you.) These are just my opinions. I wouldn't even try to prove them. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] general weak ai
Russell Wallace wrote: On 3/9/07, *Charles D Hixson* [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: Russell Wallace wrote: To test whether a program understands a story, start by having it generate an animated movie of the story. Nearly every person I know would fail THAT test. Perhaps, but what of it? I'm not trying to put forward a philosophical proof of the in-principle possibility of AI, but an agenda to improve the probability that an AI project will produce useful software, so my criterion isn't fair but likely to produce useful results. I've helped make a SHORT (30 sec.) animation. You aren't making a trivial request. It's trivial compared to AI :) You aren't requesting it of the person, you're requesting it of the AI. In other words, you are insisting that the AI demonstrate more capabilities (in a restricted domain, admittedly) than an average person before you will admit that it is intelligent. A fairer request would that it sketch a few scenes from the story (with text annotations indicating what they were supposed to represent). At this you are already requiring mastery of skills not normally attained by 4 year olds. (Well, not if the characters are supposed to be recognizable.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Languages for AGI [WAS Re: [agi] Priors and indefinite probabilities]
Chuck Esterbrook wrote: On 2/18/07, Eliezer S. Yudkowsky [EMAIL PROTECTED] wrote: Mark Waser wrote: ... I find C++ overly complex while simultaneously lacking well known productivity boosters including: * garbage collection * language level bounds checking * contracts * reflection / introspection (complete and portable) * dynamic loading (portable) * dynamic invocation Having benefited from these in other languages such as Python and C#, I'm not going back. Ever. ... Best regards, -Chuck You might check out D ( http://www.digitalmars.com/d/index.html ). Mind you, it's still in the quite early days, and missing a lot of libraries ... which means you need to construct interfaces to the C versions. Still, it answers several of your objections, and has partial answers to at least one of the others. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] Relevance of Probability
Richard Loosemore wrote: ... [ASIDE. An example of this. The system is trying to answer the question Are all ravens black?, but it does not just look to its collected data about ravens (partly represented by the vector of numbers inside the raven concept, which are vaguely related to the relevant probability), it also matters, quite crucially, that the STM contains a representation of the fact that the question is being asked by a psychologist, and that whereas the usual answer would be p(all ravens are black) = 1.0, this particular situation might be an attempt to make the subject come up with the most bizarre possible counterexamples (a genetic mutant; a raven that just had an accident with a pot of white paint, etc. etc.). In these circumstances, the numbers encoded inside concepts seem less relevant than the fact of there being a person of a particular type uttering the question.] ... Just doing my usual anarchic bit to bend the world to my unreasonable position, that's all ;-). Richard Loosemore. I would model things differently, the reactions would likely be the same, but ... One encounters an assertion All ravens are black. (in some context) One immediately hits memories of previously encountering this (or equivalent?) statements. One then notices that one hasn't encountered any ravens that aren't black. Then one creates a tentative acknowledgement Yes, all ravens are black. One evaluates the importance of an accurately correct answer (in the current context). If approximate is good enough, one sticks with this acknowledgement. If, however, it's important to be precisely accurate, one models the world, examining what features might cause a raven to not be black. If some are found, then one modifies the statement, thus: All ravens are black, except for special circumstances. One checks to see whether this suffices. If not, then one begins attaching a list of possible special circumstances, in order of generation from the list. All ravens are black, except for special circumstances, such as: they've acquired a coat of paint (or other coloring material), there might be a mutation that would change their color, etc. The significant thing here is that there are many stages where the derivation could be truncated. At each stage a check is made if it's necessary to continue. Just how precise an answer is needed? Your example of a psychologist asking the question shapes the frame of the quest for sufficiently precise, but it's always present. Rarely does one calculate a complete answer. Usually one either stops at good enough, or retrieves a appropriate answer from memory. Note that I implicitly asserted that, in this case, modeling the world was more expensive than retrieving from memory. That's because that's how I experienced it. It is, however, not always true. Also, if the answer to a question is dependent on the current context, then modeling the world may well be the only way to derive an answer. (Memory will still be used to set constraints and suggest approaches. This is because that approach is faster and more efficient that calculating such things de novo...and often more accurate.) This is related to the earlier discussion on optimality. I feel that generally minds don't even attempt optimality as normally defined, but rather search for a least cost method that's good enough. Of course, if several good enough methods are available the most nearly optimal will often be chosen. Not always though. Exploration is a part of what minds do. A lot depends on what the pressures are at the moment. One could consider this exploration as the search for a more nearly optimal method, but I'm not sure that's an accurate characterization. I rather suspect that what's happening is a getting to know the environment. Of course, one could always argue that in a larger context this is more nearly optimal...because minds have been selected to be more nearly optimal than the competition, but it's a global optimality, not the optimality in any particular problem. And, of course, the optimal organization of a mind historically depends upon the body that it's inhabiting. Thus beavers, cats, and humans will approach the problem of crossing a stream differently. Of them all, only the beaver is likely to have a mind that is tuned to a nearly optimal approach to that problem. (And its optimal approach would be of no use to a human or a cat, because of the requirement that minds match their bodies.) Is the AGI going to be disembodied? Then it will have a very different optimal organization that will a human. But a global optimization of the AGI will require that it initially be able to communicate with and understand the motivations of humans. This doesn't imply that humans will understand its motivations. Odds are they will do so quite poorly. They will probably easily model the AGI as if it were another
Re: [agi] foundations of probability theory
gts wrote: Hi Ben, On Extropy-chat, you and I and others were discussing the foundations of probability theory, in particular the philosophical controversy surrounding the so-called Principle of Indifference. Probability theory is of course relevant to AGI because of its bearing on decision theory (I assume that's why you invited me here. :) As you know, the Principle of Indifference (PI) states that if no reason exists to prefer any of n possibilities then each possibility should be assigned a probability equal to 1/n. The PI is known also as the Principle of Insufficient Reason, the name given it by classical probabilists who followed after Laplace, who took it for granted as a self-evident principle of logic. (It was John Maynard Keynes who later renamed it the Principle of Indifference.) I found a discussion of the Principle of Insufficient Reason in this book about decision theory: Choices: An Introduction to Decision Theory By Michael D. Resnik http://books.google.com/books?vid=ISBN0816614407id=4genrKNUkKcCpg=RA2-PA35lpg=RA2-PA35ots=wE4Uxk7bqEdq=principle+of+insufficient+reasonsig=PsMUy3fqcMgFha8Kyx2HLaC-EA8 This author criticizes the PI in two ways. His first is mainly philosophical: if there is no reason for assigning one set of probabilities rather than another, then there is no reason for assuming the states are equiprobable either. This is pretty much the same argument I was trying to make on ExI. His second objection is one we had not discussed: though the PI seems like a common-sense way to proceed under conditions of uncertainty, invoking it can sometimes lead to disastrous consequences for the decision-maker. I would add that while the PI might be useful in some situations as a heuristic device in programming AGI, perhaps some accounting should be made for the extra risk it entails. -gts You have a point, but perhaps not the one you think. The principle of indifference states that you should consider the probabilities of each case to be 1/n, it doesn't say anything about the relative costs of acting as if each was of equal weight. Generally if a situation is potentially more dangerous (not increased probability of harm, but probability of increased harm) the reward for not avoiding it will need to proportionately FAR greater. This means that even though the probabilities of occurrence must be deemed equal, one shouldn't consider them equally unless the cost of each event occurring is also equal. As such, indifference may be a poor name. As to whether the assumption of equal probability is valid...it seems a reasonable default position, but should be invested with very low certainty. This means that one should try to avoid any plans based on the presumption that equal probability is correct, but doesn't mean that there is a better default position. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] Project proposal: MindPixel 2
Philip Goetz wrote: On 1/17/07, Charles D Hixson [EMAIL PROTECTED] wrote: It's find to talk about making the data public domain, but that's not a good idea. Why not? Because public domain offers NO protection. If you want something close to what public domain used to provide, then the MIT license is a good choice. If you make something public domain, you are opening yourself to abusive lawsuits. (Those are always a possibility, but a license that disclaims responsibility offers *some* protection.) Public domain used to be a good choice (for some purposes), before lawsuits became quite so pernicious. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] Project proposal: MindPixel 2
Benjamin Goertzel wrote: And, importance levels need to be context-dependent, so that assigning them requires sophisticated inference in itself... The problem may not be so serious. Common sense reasoning may require only *shallow* inference chains, eg 5 applications of rules. So I'm very optimistic =) Your worries are only applicable to 100-page theorem-proving tasks, not really the concern of AGI. A) This is just not true, many commonsense inferences require significantly more than 5 applications of rules B) Even if there are only 5 applications of rules, the combinatorial explosion still exists. If there are 10 rules and 1 billlion knowledge items, then there may be up to 10 billion possibilities to consider in each inference step. So there are (10 billion)^5 possible 5-step inference trajectories, in this scenario ;-) Of course, some fairly basic pruning mechanisms can prune it down a lot, but, one is still left with a combinatorial explosion that needs to be dealt with via subtle means... Please bear in mind that we actually have a functional uncertain logical reasoning engine within the Novamente system, and have experimented with feeding in knowledge from files and doing inference on them. (Though this has been mainly for system testing, as our primary focus is on doing inference based on knowledge gained via embodied experience in the AGISim world.) The truth is that, if you have a lot of knowledge in your system's memory, you need a pretty sophisticated, context-savvy inference control mechanism to do commonsense inference. Also, temporal inference can be quite tricky, and introduces numerous options for combinatorial explosion that you may not be thinking about when looking at atemporal examples of commonsense inference. Various conclusions may hold over various time scales; various pieces of knowledge may become obsolete at various rates, etc. I imagine you will have a better sense of these issues once you have actually built an uncertain reasoning engine, fed knowledge into it, and tried to make it do interesting things I certainly think this may be a valuable exercise for you to do. However, until you have done it, I think it's kind of silly for you to be speaking so confidently about how you are so confident you can solve all the problems found by others in doing this kind of work!! I ask again, do you have some theoretical innovation that seems probably to allow you circumvent all these very familiar problems?? -- Ben Possibly this could be approached by partitioning the rule-set into small chunks of rules that work together, so that one didn't end up trying everything against everything else. These chunks of rules might well be context dependent, so that one would use different chunks at a dinner table than in a work shop. There would need to be ways to combine different chunks of rules, of course, so e.g. a restaurant table would be different from a dinner table, but would have overlapping sets of rules. (I hope I'm not just re-inventing frames...) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] Project proposal: MindPixel 2
Benjamin Goertzel wrote: Hi, Possibly this could be approached by partitioning the rule-set into small chunks of rules that work together, so that one didn't end up trying everything against everything else. These chunks of rules might well be context dependent, so that one would use different chunks at a dinner table than in a work shop. There would need to be ways to combine different chunks of rules, of course, so e.g. a restaurant table would be different from a dinner table, but would have overlapping sets of rules. (I hope I'm not just re-inventing frames...) The issue is how these contexts are learned. If context have to be programmer-supplied, then you ARE just reinventing frames Context formation is a tricky inference problem in itself -- Ben Well, my rather vague idea was to start with a very small rule set, that didn't need to be partitioned, and evolve rule-sets by statistical correlation (what tends to get used with what). As new rules are added, at some point clusters would need to separate (for efficiency). I suppose this could all be done with activation levels, but that's not the way I tend to think of it. OTOH, if the local cluster can't handle the deduction, it would need to check the most closely associated/most activated clusters to see if they could handle it. Not sure how well this would work. Clearly it has no more theoretical power than having all the rules in a large table, but I feel it would be a more efficient organization. Also, I don't have any definition of rule yet. It's not at all clear that it would be easy to translate into something a person not familiar with the details of the hardware and software would understand. (If a certain area of RAM is mapped to a video camera, reading/writing the ram will naturally mean something very different than it would mean in other contexts. Writing to it might be a request to alter the scene . (A silly way to do things, but it's for the sake of the point, not for real implementation.) I'm not at all sure that rules of the form if x do y, then check for result z (if not raise exception w) will suffice, even if you allow great flexibility as to what x, y, z, and w are interpreted as. Possibly if they could be generalized functions (with x and z limited to not causing side effects). - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] Project proposal: MindPixel 2
YKY (Yan King Yin) wrote: ... I think a project like this one requires substantial efforts, so people would need to be paid to do some of the work (programming, interface design, etc), especially if we want to build a high quality knowledgebase. If we make it free then a likely outcome is that we get a lot of noise but very few people actually contribute. I'm not an academic (left uni a couple years ago) so I can't get academic funding for this. If I can't start an AI business I'd have to entirely give up AI as a career. I hope you can understand these circumstances. YKY I can understand those circumstances, but if you expect people to contribute, you must give them something back. One thing that's cheap to give back is the work that they and others have contributed. Giving back less generally results in people not being willing to participate. Even if you claim sole rights to commercially exploit the work, you will find it much more difficult to get folk to participate. They will feel that you are stealing their work without just compensation. You raise the issue of compensation to you, and that's fair. But if you take out too much, you will cause the project to fail just as surely as if you hadn't put in the time to design the interface. If you merely make a requirement that people be a better than average contributor to be entitled to download the current results, then you will eliminate most potential competitors...and the remaining ones will be those who are also dedicating time and effort to making your project work. It's true that old versions of your work will circulate, but that should do little harm. People only participate in a public project if they feel they are getting a good return out of it. What a good return is, is subjective, but few people consider I put in a bunch of work, and they don't even mention my name to be a good return. You want to give people a return that they see as more valuable than their efforts, but which costs you a lot less than their efforts. Status in a community requires that the community exist. (At some point you'll want to give people scores depending on the amount of their work that is included in the current project...or something that will relate positively to that. This is a cheap status reward, and will boost community participation. On Slashdot I notice that just having a low numbered user ID has become a status marker of sorts. I.e., you've been a member of the community for a long time. That was a REALLY cheap status gift, but it took a long time to build to anything of value. Much quicker was the right to meta-moderate. Slightly less quick was the right to moderate. Note that these are both seen by the Slashdot community as things of worth, yet to the operator of Slashdot they were instituted as ways of cutting cost while improving quality. Also note that it took a long time for them to become worth much as status markers. You need something else to use while you're getting started.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] Project proposal: MindPixel 2
Joel Pitt wrote: ... Some comments/suggestions: * I think such a project should make the data public domain. Ignore silly ideas like giving be shares in the knowledge or whatever. It just complicates things. If the project is really strapped for cash later, then either use ad revenue or look for research funding (although I don't see much cost except for initial development of the system and web hosting). ... Making this proprietary and expecting shares to translate into cash would indeed be a silly approach. OTOH, having people's names attached to scores of some type (call them shares, or anything else) lets people feel more attached to the project. This is probably necessary for success. There also needs to be some way for the builders to interact, and a few other methods that assist the formation of a community. Newsboards, games, etc. can all be useful if structured properly to enhance the formation of a community. Perhaps only community members with scores above the median could be allowed to download the database? It's find to talk about making the data public domain, but that's not a good idea. There are arguments in favor of BSD, MIT, GPL, LGPL, etc. licenses. For this kind of activity I can see either BSD or MIT as easily defensible. (Personally I'd use LGPL, but then if I were using it, I'd want the whole application to be GPL. I might not be able to achieve it, but that's what I'd want.) Public domain wouldn't be one of the possibilities that I would consider. The Artistic license is about as close to that as I would want to come...and the MIT license is probably a better choice for those purposes. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Motivational Systems of an AI [WAS Re: [agi] RSI - What is it and how fast?]
Philip Goetz wrote: ... The disagreement here is a side-effect of postmodern thought. Matt is using evolution as the opposite of devolution, whereas Eric seems to be using it as meaning change, of any kind, via natural selection. We have difficulty because people with political agendas - notably Stephen J. Gould - have brainwashed us into believing that we must never speak of evolution being forward or backward, and that change in any direction is equally valuable. With such a viewpoint, though, it is impossible to express concern about the rising incidence of allergies, genetic diseases, etc. ... To speak of evolution as being forward or backward is to impose upon it our own preconceptions of the direction in which it *should* be changing. This seems...misguided. To claim that because all changes in the gene pool are evolution, that therefore they are all equally valuable is to conflate two (orthogonal?) assertions. Value is inherently subjective to the entity doing the evaluation. Evolution, interpreted as statistical changes in the gene pool, in inherently objective (though, of course, measurements of it may well be biased). Stephen J. Gould may well have been more of a populizer than a research scientist, but I feel that your criticisms of his presentations are unwarranted and made either in ignorance or malice. This is not a strong belief, and were evidence presented I would be willing to change it, but I've seen such assertions made before with equal lack of evidential backing, and find them distasteful. That Stephen J. Gould had some theories of how evolution works that are not universally accepted by those skilled in the field does not warrant your comments. Many who are skilled in the field find them either intriguing or reasonable. Some find them the only reasonable proposal. I can't speak for most, as I am not a professional evolutionary biologist, and don't know that many folk who are, but it would not surprise me to find that most evolutionary biologists found his arguments reasonable and unexceptional, if not convincing. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] The Singularity
Ben Goertzel wrote: ... According to my understanding of the Novamente design and artificial developmental psychology, the breakthrough from slow to fast incremental progress will occur when the AGI system reaches Piaget's formal stage of development: http://www.agiri.org/wiki/index.php/Formal_Stage At this point, the human child like intuition of the AGI system will be able to synergize with its computer like ability to do formal syntactic analysis, and some really interesting stuff will start to happen (deviating pretty far from our experience with human cognitive development). -- Ben I do, however, have some question about it being a hard takeoff. That depends largely on 1) how efficient the program is, and 2) what computer resources are available. To me it seems quite plausible that an AGI might start out as slightly less intelligent than a normal person, or even considerably less intelligent, with the limitation being due to the available computer time. Naturally, this would change fairly rapidly over time, but not exponentially so, or at least not super-exponentially so. If, however, the singularity is delayed because the programs aren't ready, or are too inefficient, then we might see a true hard-takeoff. In that case by the time the program was ready, the computer resources that it needs would already be plentifully available. This isn't impossible, if the program comes into existence in a few decades, but if the program comes into existence within the current decade, then there would be a soft-takeoff. If it comes into existence within the next half-decade then I would expect the original AGI to be sub-normal, due to lack of available resources. Naturally all of this is dependent on many different things. If Vista really does require as much of and immense retooling to more powerful computers as some predict, then programs that aren't dependent on Vista will have more resources available, as computer designs are forced to be faster and more capacious. (Wasn't Intel promising 50 cores on a single chip in a decade? If each of those cores is as capable as a current single core, then it will take far fewer computers netted together to pool the same computing capacity...for those programs so structured as to use the capacity.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] A question on the symbol-system hypothesis
BillK wrote: ... Every time someone (subconsciously) decides to do something, their brain presents a list of reasons to go ahead. The reasons against are ignored, or weighted down to be less preferred. This applies to everything from deciding to get a new job to deciding to sleep with your best friend's wife. Sometimes a case arises when you really, really want to do something that you *know* is going to end in disaster, ruined lives, ruined career, etc. and it is impossible to think of good reasons to proceed. But you still go ahead anyway, saying that maybe it won't be so bad, maybe nobody will find out, it's not all my fault anyway, and so on. ... BillK I think you've got a time inversion here. The list of reasons to go ahead is frequently, or even usually, created AFTER the action has been done. If the list is being created BEFORE the decision, the list of reasons not to go ahead isn't ignored. Both lists are weighed, a decision is made, and AFTER the decision is made the reasons decided against have their weights reduced. If, OTOH, the decision is made BEFORE the list of reasons is created, then the list doesn't *get* created until one starts trying to justify the action, and for justification obviously reasons not to have done the thing are useless...except as a layer of whitewash to prove that all eventualities were considered. For most decisions one never bothers to verbalize why it was, or was not, done. P.S.: ...and AFTER the decision is made the reasons decided against have their weights reduced. ...: This is to reinforce a consistent self-image. If, eventually, the decision turns our to have been the wrong one, then this must be revoked, and the alternative list reinforced. At which point one's self-image changes and one says things like I don't know WHY I would have done that, because the modified self image would not have decided in that way. P.P.S: THIS IS FABULATION. I'm explaining what I think happens, but I have no actual evidence of the truth of my assertions. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] A question on the symbol-system hypothesis
BillK wrote: On 12/5/06, Charles D Hixson wrote: BillK wrote: ... No time inversion intended. What I intended to say was that most (all?) decisions are made subconsciously before the conscious mind starts its reason / excuse generation process. The conscious mind pretending to weigh various reasons is just a human conceit. This feature was necessary in early evolution for survival. When danger threatened, immediate action was required. Flee or fight! No time to consider options with the new-fangled consciousness brain mechanism that evolution was developing. With the luxury of having plenty of time to reason about decisions, our consciousness can now play its reasoning games to justify what subconsciously has already been decided. NOTE: This is probably an exaggeration / simplification. ;) BillK I would say that all decisions are made subconsciously, but that the conscious mind can focus attention onto various parts of the problem and possibly affect the weighings of the factors. I would also make a distinction between the conscious mind and the verbalized elements, which are merely the story that the conscious mind is telling. (And assert that ALL of the stories that we tell ourselves are human conceits, i.e., abstractions of parts deemed significant out of a much more complex underlying process.) I've started reading What is Thought by Eric Baum. So far I'm only into the second chapter, but it seems quite promising. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Motivational Systems of an AI [WAS Re: [agi] RSI - What is it and how fast?]
James Ratcliff wrote: There is a needed distinctintion that must be made here about hunger as a goal stack motivator. We CANNOT change the hunger sensation, (short of physical manipuations, or mind-control stuff) as it is a given sensation that comes directly from the physical body. What we can change is the placement in the goal stack, or the priority position it is given. We CAN choose to put it on the bottom of our list of goals, or remove it from teh list and try and starve ourselves to death. Our body will then continuosly send the hunger signals to us, and we must decide what how to handle that signal. So in general, the Signal is there, but the goal is not, it is under our control. James Ratcliff That's an important distinction, but I would assert that although one can insert goals above a built-in goal (hunger, e.g.), one cannot remove that goal. There is a very long period when someone on a hunger strike must continually reinforce the goal of not-eating. The goal of satisfy hunger is only removed when the body decides that it is unreachable (at the moment). The goal cannot be removed by intention, it can only be overridden and suppressed. Other varieties of goal, volitionally chosen ones, can be volitionally revoked. Even in such cases habit can cause the automatic execution of tasks required to achieve the goal to be continued. I retired years ago, and although I no longer automatically get up at 5:30 each morning, I still tend to arise before 8:00. This is quite a contrast from my time in college when I would rarely arise before 9:00, and always felt I was getting up too early. It's true that with a minimal effort I can change things so that I get up a (nearly?) any particular time...but as soon as I relax it starts drifting back to early morning. Goals are important. Some are built-in, some are changeable. Habits are also important, perhaps nearly as much so. Habits are initially created to satisfy goals, but when goals change, or circumstances alter, the habits don't automatically change in synchrony. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Motivational Systems of an AI [WAS Re: [agi] RSI - What is it and how fast?]
Consider as a possible working definition: A goal is the target state of a homeostatic system. (Don't take homeostatic too literally, though.) Thus, if one sets a thermostat to 70 degrees Fahrenheit, then it's goal is to change to room temperature to be not less than 67 degrees Fahrenheit. (I'm assuming that the thermostat allows a 6 degree heat swing, heats until it senses 73 degrees, then turns off the heater until the temperature drops below 67 degrees.) Thus, the goal is the target at which a system (or subsystem) is aimed. Note that with this definition goals do not imply intelligence of more than the most very basic level. (The thermostat senses it's environment and reacts to adjust it to suit it's goals, but it has no knowledge of what it is doing or why, or even THAT it is doing it.) One could reasonably assert that the intelligence of the thermostat is, or at least has been, embodied outside the thermostat. I'm not certain that this is useful, but it's reasonable, and if you need to tie goals into intelligence, then adopt that model. James Ratcliff wrote: Can we go back to a simpler distictintion then, what are you defining Goal as? I see the goal term, as a higher level reasoning 'tool' Wherin the body is constantly sending signals to our minds, but the goals are all created consciously or semi-conscisly. Are you saying we should partition the Top-Level goals into some form of physical body - imposed goals and other types, or do you think we should leave it up to a single Constroller to interpret the signals coming from teh body and form the goals. In humans it looks to be the one way, but with AGI's it appears it would/could be another. James */Charles D Hixson [EMAIL PROTECTED]/* wrote: J... Goals are important. Some are built-in, some are changeable. Habits are also important, perhaps nearly as much so. Habits are initially created to satisfy goals, but when goals change, or circumstances alter, the habits don't automatically change in synchrony. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 ___ James Ratcliff - http://falazar.com New Torrent Site, Has TV and Movie Downloads! http://www.falazar.com/projects/Torrents/tvtorrents_show.php Everyone is raving about the all-new Yahoo! Mail beta. http://us.rd.yahoo.com/evt=45083/*http://advision.webevents.yahoo.com/mailbeta This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] A question on the symbol-system hypothesis
Mark Waser wrote: Hi Bill, ... If storage and access are the concern, your own argument says that a sufficiently enhanced human can understand anything and I am at a loss as to why an above-average human with a computer and computer skills can't be considered nearly indefinitely enhanced. The use of external aids doesn't allow one to increase the size of active ram. Usually this is no absolute barrier, though it can result in exponential slowdown. Sometimes, however, I suspect that there are problems that can't be addressed because the working memory is too small. This isn't a thing that I could prove (and probably von Neuman proved otherwise). So take exponential slowdown to be what's involved, though it might be combinatorial slowdown for some classes of problems. This may not be an absolute barrier, but it is sufficient to effectively be called one, especially given the expected lifetime of the person involved. (After one has lived a few thousand years, one might perceive this class of problems to be more tractable...but I'd bet they will be addressed sooner by other means.) Consider that we apparently have special purpose hardware for rotating visual images. Given that, there MUST be a limit to the resolution that this hardware possesses. (Well, I suspect that it rotates vectorized images, and retranslates after rotation...but SOME pixelated image is being rotated (they've watched it on PET[?] scans). This implies that anything that requires more than that much detail to handle is fudged, or just isn't handled. So the necessary enhancement would: 1) off-load the original image 2) rotate it, and 3) import the rotated image Plausibly importation could be done via a 3-D monitor, though it might take a lot of study. Exporting the original uncorrupted image, however, is beyond the current state of the art. I would argue that this is but one of a large class of problems that cannot be addressed by the current modes of enhancement. Regarding chess or Go masters -- while you couldn't point to a move and say you shouldn't have done that, I'm sure that the master could (probably in several instances) point to a move and say I wouldn't have done that and provided a better move (most often along with a variable-quality explanation of why it was a better move). ... Mark - Original Message - From: BillK [EMAIL PROTECTED] To: agi@v2.listbox.com Sent: Saturday, December 02, 2006 2:31 PM Subject: Re: [agi] A question on the symbol-system hypothesis ... - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: Motivational Systems of an AI [WAS Re: [agi] RSI - What is it and how fast?]
Mark Waser wrote: ... For me, yes, all of those things are good since they are on my list of goals *unless* the method of accomplishing them steps on a higher goal OR a collection of goals with greater total weight OR violates one of my limitations (restrictions). ... If you put every good thing on your list of goals, then you will have a VERY long list. I would propose that most of those items listed should be derived goals rather than anything primary. And that the primary goals should be rather few. I'm certain that three is too few. Probably it should be fewer than 200. The challenge is so phrasing them that they: 1) cover every needed situation 2) are short enough to be debugable They should probably be divided into two sets. One set would be a list of goals to be aimed for, and the other would be a list of filters that had to be passed. Think of these as the axioms on which the mind is being erected. Axioms need to be few and simple, it's the theorums that are derived from them that get complicated. N.B.: This is an ANALOGY. I'm not proposing a theorum prover as the model of an AI. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] A question on the symbol-system hypothesis
I don't know that I'd consider that an example of an uncomplicated goal. That seems to me much more complicated than simple responses to sensory inputs. Valuable, yes, and even vital for any significant intelligence, but definitely not at the minimal level of complexity. An example of a minimal goal might be to cause an extended period of inter-entity communication, or to find a recharging socket. Note that the second one would probably need to have a hard-coded solution available before the entity was able to start any independent explorations. This doesn't mean that as new answers were constructed the original might not decrease in significance and eventually be garbage collected. It means that it would need to be there as a pre-written answer on the tabula rasa. (I.e., the tablet can't really be blank. You need to start somewhere, even if you leave and never return.) For the first example, I was thinking of peek-a-boo. Bob Mottram wrote: Goals don't necessarily need to be complex or even explicitly defined. One goal might just be to minimise the difference between experiences (whether real or simulated) and expectations. In this way the system learns what a normal state of being is, and detect deviations. On 21/11/06, *Charles D Hixson* [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: Bob Mottram wrote: On 17/11/06, *Charles D Hixson* [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] mailto:[EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: A system understands a situation that it encounters if it predictably acts in such a way as to maximize the probability of achieving it's goals in that situation. I'd say a system understands a situation when its internal modeling of that situation closely approximates its main salient features, such that the difference between expectation and reality is minimised. What counts as salient depends upon goals. So for example I could say that I understand how to drive, even if I don't have any detailed knowledge of the workings of a car. When young animals play they're generating and tuning their models, trying to bring them in line with observations and goals. That sounds reasonable, but how are you determining the match of the internal modeling to the main salient features. I propose that you do this based on it's actions, and thus my definition. I'll admit, however, that this still leaves the problem of how to observe what it's goals are, but I hypothesize that it will be much simpler to examine the goals in the code than to examine the internal model. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] A question on the symbol-system hypothesis
OK. James Ratcliff wrote: Have to amend that to acts or replies I consider a reply an action. I'm presuming that one can monitor the internal state of the program. and it could react unpredictably depending on the humans level of understanding if it sees a nice neat answer, (like the jumping thru the window cause the door was blocked) that the human wasnt aware of, or was suprised about it would be equally good. I'm a long way from a AGI, so I'm not seriously considering superhuman understanding. That said, I proposing that you are running the system through trials. Once it has learned a trial, we say it understands the trial if it responds correctly. Correctly is defined in terms of the goals of the system rather than in terms of my goals. And this doesnt cover the opposite of what other actions can be done, and what are the consequences, that is also important. True. This doesn't cover intelligence or planning, merely understanding. And lastly this is for a situation only, we also have the more general case about understading a thing Where when it sees. or has, or is told about a thing, it understands it if, it know about general properties, and actions that can be done with, or using the thing. You are correct. I'm presuming that understanding is defined in a situation, and that it doesn't automatically transfer from one situation to another. (E.g., I understand English. Unless the accent is too strong. But I don't understand Hindi, though many English speakers do.) The main thing being we cant and arnt really defining understanding but the effect of the understanding, either in action or in a language reply. Does understanding HAVE any context free meaning? It might, but I don't feel that I could reasonably assert this. Possibly it depends on the precise definition chosen. (Consider, e.g., that one might choose to use the word meaning to refer to the context-free component of understanding. Would or would not this be a reasonable use of the language? To me this seems justifiable, but definitely not self-evident.) And it should be a level of understanding, not just a y/n. Probably, but this might depend on the complexity of the system that one was modeling. I definitely have a partial understanding of How to program an AGI. It's clearly less than 100%, and is probably greater than 1%. It may also depend on the precision with which one is speaking. To be truly precise one would doubtless need to decompose the measure along several dimensions...and it's not at all clear that the same dimensions would be appropriate in every context. But this is clearly not the appropriate place to start. So if one AI saw an apple and said, I can throw / cut / eat it, and weighted those ideas. and the second had the same list, but weighted eat as more likely, and/or knew people sometimes cut it before eating it. Then the AI would understand to a higher level. Likewise if instead, one knew you could bake an apple pie, or apples came from apple trees, he would understand more. No. That's what I'm challenging. You are relating the apple to the human world rather than to the goals of the AI. So it starts looking like a knowledge test then. What you are proposing looks like a knowledge test. That's not what I mean. Maybe we could extract simple facts from wiki, and start creating a test there, then add in more complicated things. James */Charles D Hixson [EMAIL PROTECTED]/* wrote: Ben Goertzel wrote: ... On the other hand, the notions of intelligence and understanding and so forth being bandied about on this list obviously ARE intended to capture essential aspects of the commonsense notions that share the same word with them. ... Ben Given that purpose, I propose the following definition: A system understands a situation that it encounters if it predictably acts in such a way as to maximize the probability of achieving it's goals in that situation. I'll grant that it's a bit fuzzy, but I believe that it captures the essence of the visible evidence of understanding. This doesn't say what understanding is, merely how you can recognize it. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303 ___ James Ratcliff - http://falazar.com New Torrent Site, Has TV and Movie Downloads! http://www.falazar.com/projects/Torrents/tvtorrents_show.php Sponsored Link Mortgage rates as low as 4.625% - $150,000 loan for $579 a month. Intro-*Terms https://www2.nextag.com/goto.jsp?product=10035url=%2fst.jsptm=ysearch=b_rate150ks=3968p=5035disc=yvers=722 This list
Re: [agi] A question on the symbol-system hypothesis
Ben Goertzel wrote: ... On the other hand, the notions of intelligence and understanding and so forth being bandied about on this list obviously ARE intended to capture essential aspects of the commonsense notions that share the same word with them. ... Ben Given that purpose, I propose the following definition: A system understands a situation that it encounters if it predictably acts in such a way as to maximize the probability of achieving it's goals in that situation. I'll grant that it's a bit fuzzy, but I believe that it captures the essence of the visible evidence of understanding. This doesn't say what understanding is, merely how you can recognize it. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] Natural versus formal AI interface languages
Richard Loosemore wrote: ... This is a question directed at this whole thread, about simplifying language to communicate with an AI system, so we can at least get something working, and then go from there This rationale is the very same rationale that drove researchers into Blocks World programs. Winograd and SHRDLU, etc. It was a mistake then: it is surely just as much of a mistake now. Richard Loosemore. - Not surely. It's definitely a defensible position, but I don't see any evidence that it has even a 50% probability of being correct. Also I'm not certain that SHRDLU and Blocks World were mistakes. They didn't succeed in their goals, but they remain as important markers. At each step we have limitations imposed by both our knowledge and our resources. These limits aren't constant. (P.S.: I'd throw Eliza into this same category...even though the purpose behind Eliza was different.) Think of the various approaches taken as being experiments with the user interface...since that's a large part of what they were. They are, of course, also experiments with how far one can push a given technique before encountering a combinatorial explosion. People don't seem very good at understanding that intuitively. In neural nets this same problem re-appears as saturation, the point at which as you learn new things old things become fuzzier and less certain. This may have some relevance to the way that people are continually re-writing their memories whenever they remember something. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
Re: [agi] Natural versus formal AI interface languages
John Scanlon wrote: Ben, I did read your stuff on Lojban++, and it's the sort of language I'm talking about. This kind of language lets the computer and the user meet halfway. The computer can parse the language like any other computer language, but the terms and constructions are designed for talking about objects and events in the real world -- rather than for compilation into procedural machine code. Which brings up a question -- is it better to use a language based on term or predicate logic, or one that imitates (is isomorphic to) natural languages? A formal language imitating a natural language would have the same kinds of structures that almost all natural languages have: nouns, verbs, adjectives, prepositions, etc. There must be a reason natural languages almost always follow the pattern of something carrying out some action, in some way, and if transitive, to or on something else. On the other hand, a logical language allows direct translation into formal logic, which can be used to derive all sorts of implications (not sure of the terminology here) mechanically. The problem here is that when people use a language to communicate with each other they fall into the habit of using human, rather than formal, parsings. This works between people, but would play hob with a computer's understanding (if it even had reasonable referrents for most of the terms under discussion). Also, notice one major difference between ALL human languages and computer languages: Human languages rarely use many local variables, computer languages do. Even the words that appear to be local variables in human languages are generally references, rather than variables. This is (partially) because computer languages are designed to describe processes, and human languages are quasi-serial communication protocols. Notice that thoughts are not serial, and generally not translatable into words without extreme loss of meaning. Human languages presume sufficient understanding at the other end of the communication channel to reconstruct a model of what the original thought might have been. So. Lojban++ might be a good language for humans to communicate to an AI with, but it would be a lousy language in which to implement that same AI. But even for this purpose the language needs a verifier to insure that the correct forms are being followed. Ideally such a verifier would paraphrase the statement that it was parsing and emit back to the sender either an error message, or the paraphrased sentence. Then the sender would check that the received sentence matched in meaning the sentence that was sent. (N.B.: The verifier only checks the formal properties of the language to ensure that they are followed. It had no understanding, so it can't check the meaning.) - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Natural versus formal AI interface languages
BillK wrote: On 11/1/06, Charles D Hixson wrote: So. Lojban++ might be a good language for humans to communicate to an AI with, but it would be a lousy language in which to implement that same AI. But even for this purpose the language needs a verifier to insure that the correct forms are being followed. Ideally such a verifier would paraphrase the statement that it was parsing and emit back to the sender either an error message, or the paraphrased sentence. Then the sender would check that the received sentence matched in meaning the sentence that was sent. (N.B.: The verifier only checks the formal properties of the language to ensure that they are followed. It had no understanding, so it can't check the meaning.) This discussion reminds me of a story about the United Nations assembly meetings. Normally when a representative is speaking, all the translation staff are jabbering away in tandem with the speaker. But when the German representative starts speaking they all fall silent and sit staring at him. The reason is that they are waiting for the verb to come along. :) Billk Yeah, it wouldn't be ideal for rapid interaction. But it would help people to maintain adherence to the formal rules, and to notice when they weren't. If you don't have feedback of this nature, the language will evolve different rules, more closely similar to those of natural languages. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Is a robot a Turing Machine?
Pei Wang wrote: We all know that, in a sense, every computer system (hardware plus software) can be abstractly described as a Turing machine. Can we say the same for every robot? Why? Reference to previous publications are also welcome. Pei The controller for the robot might be a Turing machine, but standard Turing machines don't include manipulators, etc. I seem to remember that even I/O on a Turing machine was an infinitely long tape (one bit wide). One can create a partial isomorphism between standard I/O devices and that tape, but nothing here connects the internals of the machine to either sensors or manipulators of the external world. All robots have SOMETHING that allows them to sense and manipulate the external world on a soft real-time basis. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Why so few AGI projects?
Joshua Fox wrote: I'd like to raise a FAQ: Why is so little AGI research and development being done? ... Thanks, Joshua What proportion of the work that is being done do you believe you are aware of? On what basis? My suspicion is that most people on the track of something new tend to be rather close about it. I'll agree that this probably slows down progress, but from an individual person or corporation's point of view it is quite sensible. For one thing, it minimizes the risks of humiliation. - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] AGI open source license
Philip Goetz wrote: ... Those companies don't make money off the software. They sell products and services. The GPL is not successful at enabling people to make money directly off software. This is critical, because it takes a large company and a large capital investment to make money selling products and services. This business model is useless to people like us, who need a way to hack out some code and make money off the code. You are right. And I can understand why in such a case the GPL might not be the right license for you to use. OTOH, if your license precludes my using your software in a GPL program, then it precludes me from using it. There are tradeoffs everywhere. Your choices are yours, and appear to be for valid to you reasons (rather than due to some misunderstanding). --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] AGI open source license
Philip Goetz wrote: On 8/30/06, Charles D Hixson [EMAIL PROTECTED] wrote: ... some snipping ... - Phil The idea with the GPL is that if you want to also sell the program commercially, you should additionally make it available under an alternate license. Some companies have been successful in this mode. (Trolltech comes to mind, and also, I believe, MySQL.) Descendants of the GPL code are required to be GPL. No restrictions are placed as to what additional licenses you might offer code for sale under that you have also offered as GPL, except for What can I get people to buy?. Commonly this is used to allow those who don't wish to agree to the terms of the GPL to purchase the right to use the code under other terms. Stipulating fees as a part of the license is probably a bad idea. Why is it a bad idea? You said it was a bad idea, but you didn't say why. It's a bad idea because licenses are relatively permanent, and prices fluctuate. This is especially true if foreign currencies become involved, but it's true over time anyway. Also, the assertion that no restrictions are placed as to what additional licenses you might offer code for sale under is wrong; you are expressly forbidden from adding additional restrictions. I can't If you own the copyright to some material, you can sell it to different people under different licenses. (Note that this requires that you own the copyrights. Not just some of them, but all of them. This is one reason this approach is infrequent.) parse the sentence saying that this is to allow those who don't wish to agree to the terms of the GPL to purchase the right to use the code under other terms - it seems to be saying that it is legal to distribute GPLed code in a non-GPL way, which it isn't. You can't distribute the GPL'd copy under non-GPL terms, but if you also bought a different license (say from TrollTech), that license might well permit you to, e.g., distribute binary only copies of a modified original. This would not be under the GPL at all. The GPL prohibits this, so you need to purchase a separate license. Actually, Trolltech requires that you do your development FROM SCRATCH under the non GPL license. It is a good idea, for these reasons: 1. The money would be paid to the people who wrote the software. Under the GPL model you're promoting, the authors get nothing. The GPL does not prohibit you from selling software. It merely prohibits you from prohibiting others from selling copies of their copy at any price they choose. 2. The GPL is unworkable. It requires that the commercial code also be released under GPL, and that the source code to everything added is released. It also requires the company to relinquish patent rights to anything in the code. This is a complete non-starter. The GPL is currently successful. Few companies are successful with their main product under the GPL license...though MySQL comes to mind, and I believe that SleepyCat is even successful distributing source code under the BSD license. Examining actual cases proves your assertions incorrect. ... --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] AGI open source license
Stephen Reed wrote: ... Rather than cash payments I have in mind a scheme similar to the pre-world wide web bulletin board system in which FTP sites had upload and download ratios. If you wished to benefit from the site by downloading, you had to maintain a certain level of contributions via file uploads. Analogously, if one seeks to benefit from using a freely available internet-based distributed AGI, then one should contribute to it, either by donating some compute cycles, or by spending some time to tutor it. Cheers. -Steve But please don't block out people behind a NAT firewall. I can't fairly download via bittorrent because it won't upload through NAT. --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] AGI open source license
Philip Goetz wrote: On 8/28/06, Stephen Reed [EMAIL PROTECTED] wrote: An assumption that some may challenge is that AGI s... source license retain these benefits yet be safe? I would rather see a license which made the software free for non-commercial use, but (unlike the GNU licenses) stipulated terms, and methods of deciding fees that would be binding on the software authors, so that a company could use the software for commercial uses, provided they paid the stipulated fees to the software authors. ... - Phil The idea with the GPL is that if you want to also sell the program commercially, you should additionally make it available under an alternate license. Some companies have been successful in this mode. (Trolltech comes to mind, and also, I believe, MySQL.) Descendants of the GPL code are required to be GPL. Descendants of code acquired under the alternate license can be whatever you choose. The limitation of this approach is that it is common for the GPL branch to out-develop the non-GPL branch...so you must develop quite actively. Also you must own the copyrights to all of the code that is used. You can't add pieces from other GPL projects. Etc. No restrictions are placed as to what additional licenses you might offer code for sale under that you have also offered as GPL, except for What can I get people to buy?. Commonly this is used to allow those who don't wish to agree to the terms of the GPL to purchase the right to use the code under other terms. Stipulating fees as a part of the license is probably a bad idea. --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] AGI open source license
Stephen Reed wrote: I would appreciate comments regarding additional constraints, if any, that should be applied to a traditional open source license to achieve a free but safe widespread distribution of software that may lead to AGI. ... My personal opinion is that the best license is the GPL. Either version 2 or 3...currently I can't choose between them (partially because version 3 is still being written). Note that may large GPL projects are quite successful. Consider, e.g., gcc. The claim that because anyone CAN change the code, anyone WILL change the code is probably fallacious. Most of those who try find that their changes are less than good. Usually those who decide to create a fork find themselves being left behind by the pace of development. So generally everyone sticks with the main tree...and perhaps submits changes that they think desirable into the project. Occasionally a fork will be successful. (X Window is no longer being developed from the XFree86 tree, e.g.) But since the license is GPL, this doesn't make any difference. How do you keep the bad guys from using it? You keep on developing. Those who fork tend to fall behind, unless they get community support. Now I'll admit that this is an idealized picture of the development process, but the outline is correct. Keeping a project going takes a good manager...one who can herd cats. It requires inspiring a degree of faith and trust in people who will be working without being paid. This means you've got to inspire them as well as get them to trust you. And you've got to articulate a vision of where the project should be headed next, roadmap is the common term, without stifling creativity. P.S.: Note that gcc has several chunks. Each language has a largely separate implementation, but each needs to generate the same kind of intermediate representation. This allows several essentially independent teams to each work separately. As to just *how* independent... consider the gdc compiler ( http://sourceforge.net/projects/dgcc ). This project is prevented by licensing constraints from having ANY direct connection to the rest of gcc. Yet it can still be integrated into gcc by an end user. --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]