Re: [agi] A 1st Step To Using Your Image-ination
> Perhaps it will start to give you a sense that words and indeed all symbols > provide an extremely limited *inventory of the world* and all its infinite > parts and behaviours. > > I welcome any impressionistic responses here, including confused questions. I agree with the above, but I think one needs to be careful about levels of description... One way to define "symbol" is in accordance with Peircean semiotics ... and in this sense, not every term, predicate or variable utilized in a logical reasoning engine is actually a "symbol" from the standpoint of the reasoning/learning process implemented by the reasoning engine Similarly, if one implements a neural net learning algorithm on a digital computer, the bits used to realize the software program are symbols from the standpoint of the programming language compiler and executor, but not from the standpoint of the neural net itself... LIke neurons, logical tokens may be used as components of complex patterned arrangements, without any individual symbolic meaning. Visual images may be represented with superhuman accuracy using logical tokens for instance. These tokens are symbolic at one level, but not visually symbolic... ben --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
[agi] A 1st Step To Using Your Image-ination
Perhaps this site will help some of you to start seeing that symbols have extremely limited powers, and something more is needed - and also give you a sense of how attitudes are changing. http://www.imageandmeaning.org/ (it's part of the Envisioning Science Program - check out the movie) also: The Initiative in Innovative Computing (IIC) http://iic.harvard.edu/ No v. coherent message behind all this stuff - just a lot of ongoing questions, which I hope will get you to start asking questions. And they're mainly talking about the need to envision *science*. What they haven't realised is what follows - the need to envision and image-ine intelligence, period. But this shows things starting to happen. And the momentum will build.(Welcome info re anything related). Perhaps it will start to give you a sense that words and indeed all symbols provide an extremely limited *inventory of the world* and all its infinite parts and behaviours. I welcome any impressionistic responses here, including confused questions. --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
RE: [agi] reasoning & knowledge.. p.s.
I agree with Pei Wang 100% on this point. Even though I find many of the comments from Mike to be interesting, I think it would be much more productive to add to the solutions and problems of creating a computer based AGI rather than trying to convince the converted that AGI on today's computers is impossible. Some problems might be solved by visual techniques but if this is important then that aspect of the problem will probably have to wait until the more general problems of object extraction from video images is further along (from Bob Mottram's comments). Most of the people on this list have quite different ideas about how an AGI should be made BUT I think there are a few things that most, if not all agree on. 1. Intelligence can be created by using computers that exist today using software. 2. Physical embodiment of the software is not essential (might be desirable) for intelligence to be created. 3. Intelligence hasn't yet been reached in anyone's AGI project. It is not possible to *prove* any AGI project to be correct until it is actually an AGI and this list won't matter much when that happens. The only way to find out if a particular AGI approach is actually a good one is to try and create it. It will be difficult to identify even the right projects when they appear because the AGI will inevitably have some capabilities far beyond a humans' and other abilities that are far less. Even if a project gets some level of intelligence with a particular approach, that doesn't mean that that approach will continue to produce even higher levels of intelligence or get to the AGI level (whatever that is). Therefore, all current AGI projects have to be fundamentally based on intuition or faith or both. No argument there, but it would seem that there is no other way to get to creating an AGI when none currently exists. It is just a waste of time to demand that someone or some group produce proof that their ideas are correct when that proof is impossible to produce until an AGI is achieved. That doesn't mean we can't debate the merits of different approaches, or demonstrate why previous attempts weren't successful. The last point being very difficult because many things could result in the failure of a project including scale, resources, etc. Just because some previous approach didn't work, doesn't necessarily mean that that approach couldn't work if some other variable was changed. I believe that a paraplegic person can still be intelligent and useful if they could just type on a keyboard and use their brain. This doesn't *prove* that human intelligence can be created without a body in the first place but I think it shows that roaming around in the world and getting firsthand knowledge from a person's senses isn't a 100% prerequisite for intelligence. I would appreciate more comments on how to achieve an AGI and less on whether a AGI on computers using software is possible or not. David Clark > -Original Message- > From: Pei Wang [mailto:[EMAIL PROTECTED] > Sent: February-14-08 5:11 PM > To: agi@v2.listbox.com > Subject: Re: [agi] reasoning & knowledge.. p.s. > > You are correct that MOST PEOPLE in AI treat observation/perception as > pure passive. As on many topics, most people in AI are probably wrong. > However, you keep making claim on "everyone", "nobody", ..., which is > almost never true. If this is your way to get people to reply your > email, it won't work on me anymore. > > There are many open problems in AI, so it is not hard to find one that > haven't been solved. If you have an idea about how to solve it, then > work on it and show us how far you can go. Just saying "Nobody has > idea about how to ..." contribute little to the field, since that > problem typically has been raised decades ago. > > Pei > --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] reasoning & knowledge.. p.s.
David said: Most of the people on this list have quite different ideas about how an AGI should be made BUT I think there are a few things that most, if not all agree on. 1. Intelligence can be created by using computers that exist today using software. 2. Physical embodiment of the software is not essential (might be desirable) for intelligence to be created. 3. Intelligence hasn't yet been reached in anyone's AGI project. I agree entirely. My comments on this list are generally grounded in my own work, or my experience at Cycorp and my intuition is that indeed today's multicore computers are sufficient to achieve intelligence. Some evidence: driverless cars, e.g. the DARPA Urban Challenge, are quite competent using modest clusters of multicore computersI estimate that a near-future 8-core cpu can achieve real-time automatic speech recognition using the Sphinx-4 softwareMy own very preliminary results on an English dialog system gives me hope that a multicore cpu can be used to robustly convert text into logic faster than a human can perform the same task. For example, my Incremental Fluid Construction Grammar parser can convert "the book is on the table" into logical statements at the rate of 400 times per second per thread. That gives me a lot of headroom when expanding the grammar rule set, adding commonsense entailed facts, and pruning alternative interpretations. -Steve Stephen L. Reed Artificial Intelligence Researcher http://texai.org/blog http://texai.org 3008 Oak Crest Ave. Austin, Texas, USA 78704 512.791.7860 Never miss a thing. Make Yahoo your home page. http://www.yahoo.com/r/hs --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
RE: [agi] A 1st Step To Using Your Image-ination
Mike, You have been pushing this anti-symbol/pro-image dichotomy for a long time. I don't understand it. Images are set, or nets, of symbols. So, if, as you say " all symbols provide an extremely limited *inventory of the world* and all its infinite parts and behaviors " then images are equally limited, since they are nothing but set or nets of symbols. Your position either doesn't make sense or is poorly stated. What you are saying is somewhat like the statement that "people don't matter in politics, institutions do." Such a statement ignores the fact that institutions are made of people. But given the human mind's ability to find that portions of a statement that makes sense (the ability that enables metaphor work), one hearing this statement might understand it as implying that people acting together are more important in politics than people acting alone. Perhaps your viewpoint is that merely considering symbols operating alone or in small numbers fails to explain many important aspects of human-like intelligence. If so, that makes sense. But that idea is shared by many people on this list. Hofstadter's Copycat and his fluid reasoning approach, which has been praised by many on this list (Pei Wang worked with Hofstadter), is based exactly on the idea of computations that involve so many individual actors that many of its processes become "liquid" -- in much the same sense that a financial market with many purchasers and sellers becomes liquid. Hofstadter's Copycat, combined both local and global influences, as well as randomness to control a synthesis of a solution to a problem. Promising research (the Serre paper I have cited so many times before) has been done on image recognition, using digital and hierarchical memory representations (both symbolic), that -- even with the trivial amount of computation resources involved compared to that of the human visual system -- provides certain types of visual recognition that out perform humans. So feel free to keep pushing the importance of computing on complex sets or nets of symbols, such as visual images -- or the complex context that builds up as one reads a good novel. But please stop attacking the use of symbols, unless you can come up with arguments much better than you have in the past. The activation of a neuron in a human brain can be viewed as a symbol, because such neurons tend to have receptive fields, when specify what patterns of synaptic or chemical inputs will activate it to varying degrees or in varying firing patterns. So apparently the human mind does pretty will with symbols. Symbols can be probabilistic. Their meanings can be context sensitive. They can represent correlations that humans haven't even explicitly considered. Even almost all so-called non-symbolic computing, such as that with neural nets, use symbols to represent their weights. All digital computers are symbolic, and one can interpret many analogue computers as being symbolic as well. So focus less on attacking symbols, and more on describing what its is about computing on large set or nets of symbols, such as those involved in images, that you think is more needed. Ed Porter -Original Message- From: Mike Tintner [mailto:[EMAIL PROTECTED] Sent: Friday, February 15, 2008 10:34 AM To: agi@v2.listbox.com Subject: [agi] A 1st Step To Using Your Image-ination Perhaps this site will help some of you to start seeing that symbols have extremely limited powers, and something more is needed - and also give you a sense of how attitudes are changing. http://www.imageandmeaning.org/ (it's part of the Envisioning Science Program - check out the movie) also: The Initiative in Innovative Computing (IIC) http://iic.harvard.edu/ No v. coherent message behind all this stuff - just a lot of ongoing questions, which I hope will get you to start asking questions. And they're mainly talking about the need to envision *science*. What they haven't realised is what follows - the need to envision and image-ine intelligence, period. But this shows things starting to happen. And the momentum will build.(Welcome info re anything related). Perhaps it will start to give you a sense that words and indeed all symbols provide an extremely limited *inventory of the world* and all its infinite parts and behaviours. I welcome any impressionistic responses here, including confused questions. --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?&; Powered by Listbox: http://www.listbox.com --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox:
Re: [agi] Goal Driven Systems and AI Dangers [WAS Re: Singularity Outcomes...]
Gah, sorry for the awfully late response. Studies aren't leaving me the energy to respond to e-mails more often than once in a blue moon... On Feb 4, 2008 8:49 PM, Richard Loosemore <[EMAIL PROTECTED]> wrote: > They would not operate at the "proposition level", so whatever > difficulties they have, they would at least be different. > > Consider [curiosity]. What this actually means is a tendency for the > system to seek pleasure in new ideas. "Seeking pleasure" is only a > colloquial term for what (in the system) would be a dimension of > constraint satisfaction (parallel, dynamic, weak-constraint > satisfaction). Imagine a system in which there are various > micro-operators hanging around, which seek to perform certain operations > on the structures that are currently active (for example, there will be > several micro-operators whose function is to take a representation such > as [the cat is sitting on the mat] and try to investigate various WHY > questions about the representation (Why is this cat sitting on this mat? > Why do cats in general like to sit on mats? Why does this cat Fluffy > always like to sit on mats? Does Fluffy like to sit on other things? > Where does the phrase 'the cat sat on the mat' come from? And so on). [cut the rest] Interesting. This sounds like it might be workable, though of course, the exact assosciations and such that the AGI develops sound hard to control. But then, that'd be the case for any real AGI system... > > Humans have lots of desires - call them goals or motivations - that > > manifest in differing degrees in different individuals, like wanting > > to be respected or wanting to have offspring. Still, excluding the > > most basic ones, they're all ones that a newborn child won't > > understand or feel before (s)he gets older. You could argue that they > > can't be inborn goals since the newborn mind doesn't have the concepts > > to represent them and because they manifest variably with different > > people (not everyone wants to have children, and there are probably > > even people who don't care about the respect of others), but still, > > wouldn't this imply that AGIs *can* be created with in-built goals? Or > > if such behavior can only be implemented with a motivational-system > > AI, how does that avoid the problem of some of the wanted final > > motivations being impossible to define in the initial state? > > I must think about this more carefully, because I am not quite sure of > the question. > > However, note that we (humans) probably do not get many drives that are > introduced long after childhood, and that the exceptions (sex, > motherhood desires, teenage rebellion) could well be sudden increases in > the power of drives that were there from the beginning. > > Ths may not have been your question, so I will put this one on hold. Well, the basic gist was this: you say that AGIs can't be constructed with built-in goals, because a "newborn" AGI doesn't yet have built up the concepts needed to represent the goal. Yet humans seem tend to have built-in (using the term a bit loosely, as all goals do not manifest in everyone) goals, despite the fact that newborn humans don't yet have built up the concepts needed to represent those goals. It is true that many of those drives seem to begin in early childhood, but it seems to me that there are still many goals that aren't activated until after infancy, such as the drive to have children. -- http://www.saunalahti.fi/~tspro1/ | http://xuenay.livejournal.com/ Organizations worth your time: http://www.singinst.org/ | http://www.crnano.org/ | http://lifeboat.com/ --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
Re: [agi] A 1st Step To Using Your Image-ination
On 15/02/2008, Ed Porter <[EMAIL PROTECTED]> wrote: > Mike, > > You have been pushing this anti-symbol/pro-image dichotomy for a long time. > I don't understand it. > > Images are set, or nets, of symbols. So, if, as you say > > > " all symbols provide an extremely limited *inventory of the > > world* and all its infinite parts and behaviors " > > then images are equally limited, since they are nothing but set or nets of > symbols. Your position either doesn't make sense or is poorly stated. I think the definition of symbols, is what is the problem is here. I tend to think of symbol (in an AI sense at least) to be about or related to something in the world. The classic idea of having symbols for cat or dog, and deducing facts from them. An image is not intrinsically about anything in the world, the optical illusions (dalmatian in spots, two faces or vase or the necker cube) show we can view an image in different ways. Mental Images aren't even necessarily made up of data "about" photon activity, they can be entirely concocted. Mike needs to clarify what he means by symbol before we start, or perhaps find or invent a less confusing word. Will Pearson --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
RE: [agi] A 1st Step To Using Your Image-ination
In response to William Pearson's message below. I agree a definition of symbol might be helpful. By symbol I just mean something that represents something other than itself. There need not be any explicit definition of what it represents. In some cases what it represents may only be defined by what causes it to have various activations, and what effects those various activations have. Often a symbol represents a set or a range of patterns, including hierarchies or chains of associations, including probabilistic associations. They can represent various degrees of match. Ed Porter -Original Message- From: William Pearson [mailto:[EMAIL PROTECTED] Sent: Friday, February 15, 2008 7:10 PM To: agi@v2.listbox.com Subject: Re: [agi] A 1st Step To Using Your Image-ination On 15/02/2008, Ed Porter <[EMAIL PROTECTED]> wrote: > Mike, > > You have been pushing this anti-symbol/pro-image dichotomy for a long time. > I don't understand it. > > Images are set, or nets, of symbols. So, if, as you say > > > " all symbols provide an extremely limited *inventory of the > > world* and all its infinite parts and behaviors " > > then images are equally limited, since they are nothing but set or nets of > symbols. Your position either doesn't make sense or is poorly stated. I think the definition of symbols, is what is the problem is here. I tend to think of symbol (in an AI sense at least) to be about or related to something in the world. The classic idea of having symbols for cat or dog, and deducing facts from them. An image is not intrinsically about anything in the world, the optical illusions (dalmatian in spots, two faces or vase or the necker cube) show we can view an image in different ways. Mental Images aren't even necessarily made up of data "about" photon activity, they can be entirely concocted. Mike needs to clarify what he means by symbol before we start, or perhaps find or invent a less confusing word. Will Pearson --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?&; Powered by Listbox: http://www.listbox.com --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com <>
Re: [agi] A 1st Step To Using Your Image-ination
RE: [agi] A 1st Step To Using Your Image-inationEd:I agree a definition of symbol might be helpful. In general there is agreement here, but some people do use the term confusingly in an all-purpose way. 1.Symbols : Abstract signs that have NO RESEMBLANCE to the signified.Word - "tree" - no resemblance to object. Numbers "1" "2" - no resemblance to the objects or number of objects described. Logic "If p then q" - Algebra- "x + y=z" No resemblance to anything being designated. Here's where it gets a little complex - there is no agreement about this next category but it's v. important 2. Graphics/ "Icons" [like computer icons rather than Peirce's]/ Image Schemas - signs that have a real if simple and sometimes distorted OUTLINE/PATTERN RESEMBLANCE to their signified objects - Computer Icons, Stick Drawings, Geometrical Graphs, Geometrical Figures, Cartoons, Maps, Traffic Signs 3.Images - much more DETAILED RESEMBLANCE - pseudo- or even actual-recordings - photographs, movies, realistic drawings, paintings, sculptures Computers are basically symbol processors - they can only deal in Graphics/Icons-as-Symbolic-Formulae and Images-as-Symbolic-Formulae. They cannot deal in whole forms directly as humans do - cannot literally handle them and reshape them and put one on top of another to see if they fit. Hence the celebrated imagery debate, with Pylyshyn, in order to safeguard current AI, trying to maintain that the human brain also, like current computers, reduces images to symbolic formulae in order to process them, whereas others like Kosslyn deny this. --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com