[agi] singularity humor
Top ten signs the singularity has arrived http://www.deanesmay.com/posts/1152629462.shtml --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Soar vs Novamente
Just some quick comments. It appears to me that perhaps the primary topic in question is an ability to generalize or abstract knowledge to varieties of situations. I would say that for the most part Soar is very good at *representing* and *using* composable (and therefore generalized) knowledge representations, but it is not so far Soar's strong suit to *create* such knowledge representations. There has been a bit of research in the past to get Soar to do inductive learning, and those efforts have currently shifted a bit to "stepping outside" the standard Soar model and integrating in capabilities for reinforcement learning and episodic learning. However, these efforts are in early stages. For the most part when we want nice generalized knowledge in Soar (which is often, when we are trying to build robust cognitive models or intelligent agents), we engineer the abstractions and knowledge representations directly into the system.One strength of Soar (in my opinion) is that it encourages "composable" knowledge representations that can rapidly "assemble themselves" (again with the proper hard-coded engineering) into wide varieties of actions or solutions to problems. So for example, rather than having 1000 different schemas for opening different kinds of doors, or one monolithic high-level schema, the typical approach in Soar would be to engineer independently the various small steps that can compose into a variety of door-opening schemas, and then layer on top of those low-level actions a hierarchy of potential situations (or partial situations) in which the various steps would be appropriate to execute. Done "correctly", this can lead to a robust reasoning system that can easily switch its behavior as the environment changes.However, there is a big caveat here. Although I claim (and believe) that Soar encourages the development of such robust models, it does not *require* you to represent your knowledge that way. It is certainly easy to build brittle systems in Soar, containing knowledge that is not abstracted well. An engineer has to do the work of finding the right abstractions, which it sounds to me like where some of the focus is in Novamente. Once you have some reasonable abstractions, though, Soar provides a good engine for representing the knowledge in modular and efficient ways.Randy JonesBen Goertzel [EMAIL PROTECTED] wrote: One of the key ideas underlying the NM design is to fully integratethe top-down (logical problem solving and reasoning) based approachwith the bottom-up (unsupervised, reinforcement-learning-basedstatistical pattern recognition) based approach.SOAR basically lies firmly in the former camp...-- BenOn 7/12/06, Yan King Yin <[EMAIL PROTECTED]> wrote: (From a former Soar researcher) [...] Generally, the bottom-up pattern based systems do better at noisy pattern recognition problems (perception problems like recognizing letters in scanned OCR text or building complex perception-action graphs where the decisions are largely probabilistic like playing backgammon or assigning labels to chemical molecules). Top-down reasoning systems like Soar generally do better at higher level reasoning problems. Selecting the correct formation and movements for a squad of troops when clearing a building, or receiving English instructions from a human operator to guide a robot through a burning building. [...] Doug From what I read, Soar also deals with (or has provisos to deal with) sensory processing, otherwise it wouldn't be the "unified cognitive architecture" as Allen Newell has intended it to be. The difference in emphasis between Novamente on perceptual learning and Soar on top-down reasoning, may be real but ideally it should not be accepted prima facie . IMO these 2 emphases should be integrated seamlessly. YKY To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]---To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]Thank YouJames Ratcliffhttp://falazar.com Do you Yahoo!? Get on board. You're invited to try the new Yahoo! Mail Beta. To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Computing Intelligence? How too? ................. ping
Danny, I just read an interesting article that goes through a more formal proof of intelligence "A formal measure of Machine Intelligence"which I may have gotten from a link from this group or another.http://www.vetta.org/documents/ui_benelearn.pdfJames Ratcliff[EMAIL PROTECTED] wrote: Does anyone know where I might find information on Fitness algorithms? How does one determine the level of Intelligence of any given AI System configuration? What are the best methods of computing algorithm fitness or intelligence? Dan GoeFrom : Samantha Atkins <[EMAIL PROTECTED]>To : agi@v2.listbox.comSubject : [agi] pingDate : Wed, 5 Jul 2006 12:47:58 -0700 No mail seen since 6/30. Testing. --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]---To unsubscribe, change your address, or temporarily deactivate your subscription,please go to http://v2.listbox.com/member/[EMAIL PROTECTED]Thank YouJames Ratcliffhttp://falazar.com Sneak preview the all-new Yahoo.com. It's not radically different. Just radically better. To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Thanks James.. [agi] Computing Intelligence? How too? ................. ping
James Many thanks for the link on Computing Intelligence. Dan Goe From : James Ratcliff [EMAIL PROTECTED] To : agi@v2.listbox.com Subject : Re: [agi] Computing Intelligence? How too? . ping Date : Thu, 13 Jul 2006 07:44:26 -0700 (PDT) Danny, I just read an interesting article that goes through a more formal proof of intelligence A formal measure of Machine Intelligence which I may have gotten from a link from this group or another. http://www.vetta.org/documents/ui_benelearn.pdf James Ratcliff [EMAIL PROTECTED] wrote: Does anyone know where I might find information on Fitness algorithms? How does one determine the level of Intelligence of any given AI System configuration? What are the best methods of computing algorithm fitness or intelligence? Dan Goe From : Samantha Atkins To : agi@v2.listbox.com Subject : [agi] ping Date : Wed, 5 Jul 2006 12:47:58 -0700 No mail seen since 6/30. Testing. --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] Thank You James Ratcliff http://falazar.com - Sneak preview the all-new Yahoo.com. It's not radically different. Just radically better. --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
[agi] Thanks James... Computing Intelligence? How too? ................. ping
James, Many thanks for the link on Computing Intelligence. Dan Goe From : James Ratcliff [EMAIL PROTECTED] To : agi@v2.listbox.com Subject : Re: [agi] Computing Intelligence? How too? . ping Date : Thu, 13 Jul 2006 07:44:26 -0700 (PDT) Danny, I just read an interesting article that goes through a more formal proof of intelligence A formal measure of Machine Intelligence which I may have gotten from a link from this group or another. http://www.vetta.org/documents/ui_benelearn.pdf James Ratcliff [EMAIL PROTECTED] wrote: Does anyone know where I might find information on Fitness algorithms? How does one determine the level of Intelligence of any given AI System configuration? What are the best methods of computing algorithm fitness or intelligence? Dan Goe From : Samantha Atkins To : agi@v2.listbox.com Subject : [agi] ping Date : Wed, 5 Jul 2006 12:47:58 -0700 No mail seen since 6/30. Testing. --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] Thank You James Ratcliff http://falazar.com - Sneak preview the all-new Yahoo.com. It's not radically different. Just radically better. --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Soar vs Novamente
Soar, like other cognitive architectures (such as ACT-R), is not designed to directly deal with domain problems. Instead, it is a high-level platform on which a program can be built for a specific problem. On the contrary, Novamente, like other AGI systems (such as NARS), is designed to directly deal with domain problems. To work well, usually it needs to be trained with domain-specific knowledge, but such a training process is fundamentally different from a programming process. To me, many other differences, such as the role of learning, follow from the above difference between program to work and learn to work. The current issue of AI Magazine (http://www.aaai.org/Library/Magazine/vol27.php#Summer) is highly relevant to this discussion. Especially, the articiles by Langley, Cassimatis, and JonesWray provide good introductions and discussions about cognitive architectures. Pei On 7/13/06, Ben Goertzel [EMAIL PROTECTED] wrote: Thanks, Randy. This is very well put. Yes, one of the key things missing in rule and logic based AI systems like SOAR is the learning of new representations to match new situations and problems. Interestingly, this is also one of the key things missing in evolutionary learning as conventionally implemented. My colleague Moshe Looks has been working on a modified approach to evolutionary learning that involves automatically learning new representations for new problems; it is called MOSES and is being written for integration into Novamente as well as for standalone use. Some information on MOSES is here if you're curious: http://metacog.org/doc.html -- Ben On 7/13/06, James Ratcliff [EMAIL PROTECTED] wrote: Just some quick comments. It appears to me that perhaps the primary topic in question is an ability to generalize or abstract knowledge to varieties of situations. I would say that for the most part Soar is very good at *representing* and *using* composable (and therefore generalized) knowledge representations, but it is not so far Soar's strong suit to *create* such knowledge representations. There has been a bit of research in the past to get Soar to do inductive learning, and those efforts have currently shifted a bit to stepping outside the standard Soar model and integrating in capabilities for reinforcement learning and episodic learning. However, these efforts are in early stages. For the most part when we want nice generalized knowledge in Soar (which is often, when we are trying to build robust cognitive models or intelligent agents), we engineer the abstractions and knowledge representations directly into the system. One strength of Soar (in my opinion) is that it encourages composable knowledge representations that can rapidly assemble themselves (again with the proper hard-coded engineering) into wide varieties of actions or solutions to problems. So for example, rather than having 1000 different schemas for opening different kinds of doors, or one monolithic high-level schema, the typical approach in Soar would be to engineer independently the various small steps that can compose into a variety of door-opening schemas, and then layer on top of those low-level actions a hierarchy of potential situations (or partial situations) in which the various steps would be appropriate to execute. Done correctly, this can lead to a robust reasoning system that can easily switch its behavior as the environment changes. However, there is a big caveat here. Although I claim (and believe) that Soar encourages the development of such robust models, it does not *require* you to represent your knowledge that way. It is certainly easy to build brittle systems in Soar, containing knowledge that is not abstracted well. An engineer has to do the work of finding the right abstractions, which it sounds to me like where some of the focus is in Novamente. Once you have some reasonable abstractions, though, Soar provides a good engine for representing the knowledge in modular and efficient ways. Randy Jones Ben Goertzel [EMAIL PROTECTED] wrote: One of the key ideas underlying the NM design is to fully integrate the top-down (logical problem solving and reasoning) based approach with the bottom-up (unsupervised, reinforcement-learning-based statistical pattern recognition) based approach. SOAR basically lies firmly in the former camp... -- Ben On 7/12/06, Yan King Yin wrote: (From a former Soar researcher) [...] Generally, the bottom-up pattern based systems do better at noisy pattern recognition problems (perception problems like recognizing letters in scanned OCR text or building complex perception-action graphs where the decisions are largely probabilistic like playing backgammon or assigning labels to chemical molecules). Top-down reasoning systems like Soar generally do better at higher level reasoning problems. Selecting the correct formation and movements for a squad of troops when clearing a
Re: [agi] Processing speed for core intelligence in human brain
My personal guesstimate is that what are commonly considered the higher order cognitive functions useway less than 1% of the total power estimated for the brain (and also, that the brain does them very inefficiently so a better implementation would use even less power). On the other hand, I also believe that *everything* "higher order" in the brainruns on top of and requires a massive amount of parallel pattern-matching power -- and that this is going to be the final limiting factor on reproducing human-level intelligence (and that we are probably at least ten years from the necessary processing power and architecture and algorithms for this). - Original Message - From: Joshua Fox To: agi@v2.listbox.com Sent: Thursday, July 13, 2006 1:55 PM Subject: **SPAM** [agi] Processing speed for core intelligence in human brain Greetings, I am new to the list. I hope that the following question adds something of value. Estimates for the total processing speed of intelligence in the human brain are often used as crude guides to understanding the timeline towards human-equivalent intelligence. Would someone venture to guesstimate -- even within a couple of orders of magnitude -- the total processing speed of higher order cognitive functions, in contrast to lower-order functions like sensing and actuation.(Use any definition of "higher" and "lower" order which seems reasonable to you.) I appreciate the problems with estimating human-equivalent intelligence based on raw speed, and I recognize that tightly integrated lower-order functionality may be essential to full generalintelligence. Nonetheless, it would be fascinating to learn, e.g., that the "core" of human intelligence use only 1% of the total power estimated for the brain. That would suggest that if lower order functions can be "outsourced" to the many projects now working on them, and offloaded at runtime to remote systems, then human-order raw power may be closer than we thought. Joshua To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Processing speed for core intelligence in human brain
Joshua Fox wrote: Greetings, I am new to the list. I hope that the following question adds something of value. Estimates for the total processing speed of intelligence in the human brain are often used as crude guides to understanding the timeline towards human-equivalent intelligence. Would someone venture to guesstimate -- even within a couple of orders of magnitude -- the total processing speed of higher order cognitive functions, in contrast to lower-order functions like sensing and actuation. (Use any definition of higher and lower order which seems reasonable to you.) I appreciate the problems with estimating human-equivalent intelligence based on raw speed, and I recognize that tightly integrated lower-order functionality may be essential to full general intelligence. Nonetheless, it would be fascinating to learn, e.g., that the core of human intelligence use only 1% of the total power estimated for the brain. That would suggest that /if/ lower order functions can be outsourced to the many projects now working on them, and offloaded at runtime to remote systems, then human-order raw power may be closer than we thought. Joshua Joshua, I recently addressed a similar issue on the SL4 list, so here is an expanded version of my calculation for what I think is involved in higher order processing. My thoughts were geared towards estimating when the hardware would be available. (Answer: yesterday.) 1) Quick Introduction The basis for these calculations is the idea that the human cognitive system does all of its real work by keeping a set of elements simultaneously active and allowing them to constrain one another. Simple enough idea. Basis of neural nets, actors, etc. Then, starting with this idea, I use the fact that the brain is organized into cortical columns, and I would (cautiously) hypothesize that these could be implementing a grid of cells on which these elements can live, when they are active. This allows us to start talking about possible numbers for the simultaneously active elements and their operating timescale. Finally, notice that a good chunk of the cortical column real estate is probably devoted to visual processing. Now, some of this would not just be doing data driven processing (which would come under the heading of peripheral work, which we want to keep out of the calculation) but interactive processing that includes top-down constraints. Difficult to say how much of this visual processing really counts as higher order thought, but my guess would be that some fraction of it is not. 2) The Calculation Itself Approximate number of cortical columns: 1,000,000. If each of these is hosting a single concept, but they are providing a facility for moving the concept from one column to the next in real time, to allow concepts to make transient connections to near neighbors, then most of them may be just available for liquidity purposes (imagine a chinese puzzle on a large scale... more empty blocks means more potential for the blocks to move around, and hence greater liquidity). So, number of simultaneously active processes will be much less than 1,000,000. My use of the cortical coumn idea is really just meant as an upper bound: I am not committed to this interpretation of what the columns are doing. Second datum to use: the sensorium (the sum total of what is actively involved in our current representation of the state of the world and the content of our abstract thoughts) is likely to contain much less than 1,000,000 simultaneously active concepts. Why? Mostly because the contents of a good sized encyclopaedia would involve less than a million concepts, and we barely have enough words in our language for that many distinct, nameable concepts. It is hard to believe that we keep anything like that many concepts active at once. Using the above two factors, we could hazard a guess at perhaps as few as 10,000 simultaneously active high-level concepts, not a million. My gut feeling is that this is a conservative estimate (i.e. too high). Further suppose that the function of concepts, when active, is to engage in relatively simple interactions with neighbors in order to carry out multiple simultaneous relaxation along several dimensions. When the concepts are not active they have to go through different sorts of calculations (debriefing after an episode of being used), and when they are being activated they have to (effectively) travel from their home column to where they are needed. Considering these other computations together we notice that the cortical column may implement multiple functions that do not need to be simultaneously active. Now, all of the above functions are consistent with the complexity and layout of the columns. Notice that what is actually being computed is relatively simple, but because of the nature of the column wiring the functions take a good deal of wiring to
Re: [agi] Flow charts? Source Code? .. Computing Intelligence? How too? ................. ping
James,Currently I'm writing a much longer paper (about 40 pages) on intelligencemeasurement. A draft version of this will be ready in about a month whichI hope to circulate around a bit for comments and criticism. There is also another guy who has recently come to my attention who is doing verysimilar stuff. He has a 50 page paper on formal measures of machineintelligence that should be coming out in coming months.I'll make a post here when either of these papers becomes available. Shane To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Flow charts? Source Code? .. Computing Intelligence? How too? ................. ping
On 7/13/06, Pei Wang [EMAIL PROTECTED] wrote: Shane,Do you mean Warren Smith?Yes.Shane To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Flow charts? Source Code? .. Computing Intelligence? How too? ................. ping
Shane, Thanks, I would appreciate that greatly.On the topic of measuring intelligence, what do you think about the actual structure of comparison of some of today's AI systems. I would like to see someone come up with and get support for a general fairly widespread set of test s for general AI other than the turing test. I have recently been working with some testing stuff with the KM from UT. It and two other systems took and passed a AP exam for chemistry, which, though limited, is an impressive feat itself.James RatcliffShane Legg [EMAIL PROTECTED] wrote: James,Currently I'm writing a much longer paper (about 40 pages) on intelligencemeasurement. A draft version of this will be ready in about a month whichI hope to circulate around a bit for comments and criticism. There is also another guy who has recently come to my attention who is doing verysimilar stuff. He has a 50 page paper on formal measures of machineintelligence that should be coming out in coming months.I'll make a post here when either of these papers becomes available. Shane To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] Thank YouJames Ratcliffhttp://falazar.com How low will we go? Check out Yahoo! Messengers low PC-to-Phone call rates. To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] Flow charts? Source Code? .. Computing Intelligence? How too? ................. ping
I think that public learning/training of an AGI would be a terrible disaster... Look at what happened with OpenMind and MindPixel These projects allowed the public to upload knowledge into them, which resulted in a lot of knowledge of the general nature Jennifer Lopez got a nice butt, etc. Jason Hutchens once showed me two versions of his statistical learning based conversation system, MegaHal. One was trained by him, the other by random web-surfers. The former displayed some occasional apparent intelligence, the latter constantly spewed amusing but eventually boring junk about penises and such. I had the idea once to teach an AI system in Lojban, and then let random Lojban speakers over the Web interact with it to teach it. This might work, because the barrier to entry is so high. Anyone who has bothered to learn Lojban is probably a serious nerd and wouldn't feel like filling the AI's mind with a bunch of junk. Of course, I haven't bothered to learn Lojban well yet, though ;-( ... -- Ben On 7/13/06, James Ratcliff [EMAIL PROTECTED] wrote: Ben Goertzel [EMAIL PROTECTED] wrote: While AIXI is all a bit pie in the sky, mathematical philosophy if you like, I think the above does however highlight something of practical importance: Even if your AI is incomputably super powerful, like AIXI, the training and education of the AI is still really important. Very few people spend time thinking about how to teach and train a baby AI. I think this is a greatly ignored aspect of AI. Agree, but there is a reason: before a baby AI is actually built, not to much can be said about its education. For example, assume both AIXI and NARS are successfully built, they will need to be educated in quite different ways (though there will be some similarity), given the different design. I'll worry about education after the details of the system are relatively stable. Ben Goertzel [EMAIL PROTECTED] wrote: While AIXI is all a bit pie in the sky, mathematical philosophy if you like, I think the above does however highlight something of practical importance: Even if your AI is incomputably super powerful, like AIXI, the training and education of the AI is still really important. Very few people spend time thinking about how to teach and train a baby AI. I think this is a greatly ignored aspect of AI. Agree, but there is a reason: before a baby AI is actually built, not to much can be said about its education. For example, assume both AIXI and NARS are successfully built, they will need to be educated in quite different ways (though there will be some similarity), given the different design. I'll worry about education after the details of the system are relatively stable. Pei, I think you are right that the process of education and mental development is going to be different for different types of AGI systems. However, I don't think it has to be dramatically different for each very specific AGI design. And I don't think one has to wait till one has a working AGI to put serious analysis into its psychological development and instruction. In the context of Novamente, I have put a lot of thought into how mental development should occur for AGI systems that are -- heavily based on uncertain inference -- embodied in a real or simulated world where they get to interact with other agents Novamente falls into this category, but so do other AGI designs. A few of my and Stephan Bugaj's thoughts on this are described here: http://www.agiri.org/forum/index.php?showtopic=158 and here: http://www.novamente.net/engine/ (see Stage of Cognitive Development...) I have a whole lot of informal notes written down on AGI Developmental Psychology, extending the general ideas in this presentation/paper, and will probably write them up as a manuscript one day... -- Ben --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] Pei, I think you are right that the process of education and mental development is going to be different for different types of AGI systems. However, I don't think it has to be dramatically different for each very specific AGI design. And I don't think one has to wait till one has a working AGI to put serious analysis into its psychological development and instruction. In the context of Novamente, I have put a lot of thought into how mental development should occur for AGI systems that are -- heavily based on uncertain inference -- embodied in a real or simulated world where they get to interact with other agents Novamente falls into this category, but so do other AGI designs. A few of my and Stephan Bugaj's thoughts on this are described here: http://www.agiri.org/forum/index.php?showtopic=158 and here: http://www.novamente.net/engine/ (see Stage of Cognitive Development...) I have a whole lot of informal notes written down on AGI Developmental Psychology,
Re: [agi] Flow charts? Source Code? .. Computing Intelligence? How too? ................. ping
Ben, Yes, but OpenMind did get quite a bit of usable information into it as well, and mainly they learned a lot about the process. I believe, and they are looking at as well, different ways of grading the participants themselves, so the obviously juvienile ones could be graded down and out of the system. Likewise the processes themselves could be graded as to functionality and correctness, with the ability of a user to look at multiple task processes like "Pick up the Ball" and vote on ones that are more functional.At the very least, I would like to open it up to a number of people, and that would speed along the creation of many processes faster than I alone could ever do.James RatcliffBen Goertzel [EMAIL PROTECTED] wrote: I think that public learning/training of an AGI would be a terrible disaster...Look at what happened with OpenMind and MindPixel These projectsallowed the public to upload knowledge into them, which resulted in alot of knowledge of the general nature "Jennifer Lopez got a nicebutt", etc.Jason Hutchens once showed me two versions of his statistical learningbased conversation system, MegaHal. One was trained by him, the otherby random web-surfers. The former displayed some occasional apparentintelligence, the latter constantly spewed amusing but eventuallyboring junk about penises and such.I had the idea once to teach an AI system in Lojban, and then letrandom Lojban speakers over the Web interact with it to teach it.This might work, because the barrier to entry is so high. Anyone whohas bothered to learn Lojban is probably a serious nerd and wouldn'tfeel like filling the AI's mind with a bunch of junk. Of course, Ihaven't bothered to learn Lojban well yet, though ;-( ...-- BenOn 7/13/06, James Ratcliff <[EMAIL PROTECTED]> wrote: Ben Goertzel <[EMAIL PROTECTED]> wrote:While AIXI is all a bit pie in the sky, "mathematical philosophy" if you like, I think the above does however highlight something of practical importance: Even if your AI is incomputably super powerful, like AIXI, the training and education of the AI is still really important. Very few people spend time thinking about how to teach and train a baby AI. I think this is a greatly ignored aspect of AI. Agree, but there is a reason: before a "baby AI" is actually built, not to much can be said about its education. For example, assume both AIXI and NARS are successfully built, they will need to be educated in quite different ways (though there will be some similarity), given the different design. I'll worry about education after the details of the system are relatively stable. Ben Goertzel <[EMAIL PROTECTED]> wrote:While AIXI is all a bit pie in the sky, "mathematical philosophy" if you like, I think the above does however highlight something of practical importance: Even if your AI is incomputably super powerful, like AIXI, the training and education of the AI is still really important. Very few people spend time thinking about how to teach and train a baby AI. I think this is a greatly ignored aspect of AI. Agree, but there is a reason: before a "baby AI" is actually built, not to much can be said about its education. For example, assume both AIXI and NARS are successfully built, they will need to be educated in quite different ways (though there will be some similarity), given the different design. I'll worry about education after the details of the system are relatively stable. Pei, I think you are right that the process of education and mental development is going to be different for different types of AGI systems. However, I don't think it has to be dramatically different for each very specific AGI design. And I don't think one has to wait till one has a working AGI to put serious analysis into its psychological development and instruction. In the context of Novamente, I have put a lot of thought into how mental development should occur for AGI systems that are -- heavily based on uncertain inference -- embodied in a real or simulated world where they get to interact with other agents Novamente falls into this category, but so do other AGI designs. A few of my and Stephan Bugaj's thoughts on this are described here: http://www.agiri.org/forum/index.php?showtopic=158 and here: http://www.novamente.net/engine/ (see "Stage of Cognitive Development...") I have a whole lot of informal notes written down on AGI Developmental Psychology, extending the general ideas in this presentation/paper, and will probably write them up as a manuscript one day... -- Ben --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] Pei, I think you are right that the process of education and mental development is going to be different for different types of AGI systems. However, I don't think it has to be dramatically different for each very specific AGI design. And I don't think one has to wait
Re: [agi] Flow charts? Source Code? .. Computing Intelligence? How too? ................. ping
I agree that using the Net to recruit a team of volunteer AGI teachers would be a good idea. But opening the process up to random web-surfers is, IMO, asking for trouble...! -- Ben On 7/13/06, James Ratcliff [EMAIL PROTECTED] wrote: Ben, Yes, but OpenMind did get quite a bit of usable information into it as well, and mainly they learned a lot about the process. I believe, and they are looking at as well, different ways of grading the participants themselves, so the obviously juvienile ones could be graded down and out of the system. Likewise the processes themselves could be graded as to functionality and correctness, with the ability of a user to look at multiple task processes like Pick up the Ball and vote on ones that are more functional. At the very least, I would like to open it up to a number of people, and that would speed along the creation of many processes faster than I alone could ever do. James Ratcliff Ben Goertzel [EMAIL PROTECTED] wrote: I think that public learning/training of an AGI would be a terrible disaster... Look at what happened with OpenMind and MindPixel These projects allowed the public to upload knowledge into them, which resulted in a lot of knowledge of the general nature Jennifer Lopez got a nice butt, etc. Jason Hutchens once showed me two versions of his statistical learning based conversation system, MegaHal. One was trained by him, the other by random web-surfers. The former displayed some occasional apparent intelligence, the latter constantly spewed amusing but eventually boring junk about penises and such. I had the idea once to teach an AI system in Lojban, and then let random Lojban speakers over the Web interact with it to teach it. This might work, because the barrier to entry is so high. Anyone who has bothered to learn Lojban is probably a serious nerd and wouldn't feel like filling the AI's mind with a bunch of junk. Of course, I haven't bothered to learn Lojban well yet, though ;-( ... -- Ben On 7/13/06, James Ratcliff wrote: Ben Goertzel wrote: While AIXI is all a bit pie in the sky, mathematical philosophy if you like, I think the above does however highlight something of practical importance: Even if your AI is incomputably super powerful, like AIXI, the training and education of the AI is still really important. Very few people spend time thinking about how to teach and train a baby AI. I think this is a greatly ignored aspect of AI. Agree, but there is a reason: before a baby AI is actually built, not to much can be said about its education. For example, assume both AIXI and NARS are successfully built, they will need to be educated in quite different ways (though there will be some similarity), given the different design. I'll worry about education after the details of the system are relatively stable. Ben Goertzel wrote: While AIXI is all a bit pie in the sky, mathematical philosophy if you like, I think the above does however highlight something of practical importance: Even if your AI is incomputably super powerful, like AIXI, the training and education of the AI is still really important. Very few people spend time thinking about how to teach and train a baby AI. I think this is a greatly ignored aspect of AI. Agree, but there is a reason: before a baby AI is actually built, not to much can be said about its education. For example, assume both AIXI and NARS are successfully built, they will need to be educated in quite different ways (though there will be some similarity), given the different design. I'll worry about education after the details of the system are relatively stable. Pei, I think you are right that the process of education and mental development is going to be different for different types of AGI systems. However, I don't think it has to be dramatically different for each very specific AGI design. And I don't think one has to wait till one has a working AGI to put serious analysis into its psychological development and instruction. In the context of Novamente, I have put a lot of thought into how mental development should occur for AGI systems that are -- heavily based on uncertain inference -- embodied in a real or simulated world where they get to interact with other agents Novamente falls into this category, but so do other AGI designs. A few of my and Stephan Bugaj's thoughts on this are described here: http://www.agiri.org/forum/index.php?showtopic=158 and here: http://www.novamente.net/engine/ (see Stage of Cognitive Development...) I have a whole lot of informal notes written down on AGI Developmental Psychology, extending the general ideas in this presentation/paper, and will probably write them up as a manuscript one day... -- Ben --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL
Re: [agi] Flow charts? Source Code? .. Computing Intelligence? How too? ................. ping
Ben, Though Piaget is my favorite psychologist, I don't think his theory on Developmental Psychology applies to AI to the extent you suggested. One major reason is: in a human baby, the mental learning process in the mind and the biological developing process in the brain happen together, while in AI the former will occur within a mostly fixed hardware system. Also, an AI system doesn't have to first develop capabilities responsible for the survival of a human baby. As a result, for example, Novamente can do some abstract inference (a formal stage activity) before being able to recognize complicated patterns (an infantile stage activity). Of course, certain general principles of education will remain, such as to teach simple topics before difficult ones, to combine lectures with questions and exercises, to explain abstract materials with concrete examples, and so on, but I don't think we can get too much details with confidence. As for AIXI, since its input comes from a finite perception space and a real-number reward space, its output is selected from a fixed action space, and for a given history (past input and output) there is a fixed (though unknown) probability for each possible input to occur, the best training strategy will be very different from the case of Novamente, which is not based on such assumptions. Given the different research goals and assumptions about the interaction between the system and the environment, different AGI systems will have very different training/educating strategies, which are similar to each other only in a very vague sense. Furthermore, since all the systems are far from mature, any design change will require corresponding change in training. On the contrary, we cannot decide a training process first, then design the system accordingly. For these reasons, I'd rather not to spend too much time on training now, though I fully agree that it will become a major issue in the future. Pei On 7/13/06, Ben Goertzel [EMAIL PROTECTED] wrote: Pei, That is actually not correct... I would teach a baby AIXI about the same way I would teach a baby Novamente, but I assume the former would learn a lot faster... so the various stages of instruction would be passed through a lot more quickly Furthermore, I expect that the same cognitive structures that would develop within a Novamente during its learning process, would also develop within an AIXI during its learning process -- though in the AIXI these cognitive structures would exist within the currently active program being used to choose behaviors (due to its being chosen as optimal during AIXI's program space search). Please note that both AIXI and Novamente are explicitly based on uncertain probabilistic inference, so that in spite of the significant differences between the two (e.g. the latter can run on feasible computational infrastructure, and is much more complicated due to the need to fulfill this requirement), there is also a significant commonality. -- Ben On 7/13/06, Pei Wang [EMAIL PROTECTED] wrote: Ben, For example, I guess most of your ideas about how to train Novamente cannot be applied to AIXI. ;-) Pei Pei, I think you are right that the process of education and mental development is going to be different for different types of AGI systems. However, I don't think it has to be dramatically different for each very specific AGI design. And I don't think one has to wait till one has a working AGI to put serious analysis into its psychological development and instruction. In the context of Novamente, I have put a lot of thought into how mental development should occur for AGI systems that are -- heavily based on uncertain inference -- embodied in a real or simulated world where they get to interact with other agents Novamente falls into this category, but so do other AGI designs. A few of my and Stephan Bugaj's thoughts on this are described here: http://www.agiri.org/forum/index.php?showtopic=158 and here: http://www.novamente.net/engine/ (see Stage of Cognitive Development...) I have a whole lot of informal notes written down on AGI Developmental Psychology, extending the general ideas in this presentation/paper, and will probably write them up as a manuscript one day... -- Ben --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
Re: [agi] singularity humor
I think this one was the granddaddy: http://yudkowsky.net/humor/signs-singularity.txt -- Eliezer S. Yudkowsky http://singinst.org/ Research Fellow, Singularity Institute for Artificial Intelligence --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]