[agi] Guessing robots
article in NS about the Purdue guessing robot navigators... http://www.newscientisttech.com/article/dn11805-guessing-robots-navigate-faster.html I think I get a toljaso on this one --- if the architecture were composed of modules that did CBR, each in its own language, from the very start, this technique would have fallen out automatically. 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/?member_id=231415user_secret=fabd7936
Re: [agi] Guessing robots
I don't know what algorithms are being referred to in this article - perhaps a type of monte carlo localization. Does anyone have more direct references? Also it's only in 2D, which is normal for laser based mapping. It's unlikely that we'll see products based on this sort of technology appearing any time soon, unless someone comes out with a cheap scanning laser rangefinder (currently costing around $2000 each) or alternative types of sensor are used, such as stereo vision. On 10/05/07, J Storrs Hall, PhD [EMAIL PROTECTED] wrote: article in NS about the Purdue guessing robot navigators... http://www.newscientisttech.com/article/dn11805-guessing-robots-navigate-faster.html I think I get a toljaso on this one --- if the architecture were composed of modules that did CBR, each in its own language, from the very start, this technique would have fallen out automatically. 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/?member_id=231415user_secret=fabd7936
Re: [agi] Guessing robots
You could probably buy 10 cheap webcams and put them all around the robot and get some vision algorithms to turn them into 3D scenes, which are avoided/mapped? This seems like a pretty well understood and constrained problem. It also sounds like a lot of robot perception work on object permanence, which from what I understand is getting a noisy perceptual input to conform to a unique symbolic representation of the object (e.g. this is the same T-shaped-part-of-the-hallway that I was at 5 minutes ago.) They mention the SLAM algorithm as being similar, and it doesn't seem to have any much greater abilities to accomplish goals for example. It basically sounds like a better inference algorithm. Bo On Thu, 10 May 2007, Bob Mottram wrote: ) I don't know what algorithms are being referred to in this article - ) perhaps a type of monte carlo localization. Does anyone have more ) direct references? Also it's only in 2D, which is normal for laser ) based mapping. ) ) It's unlikely that we'll see products based on this sort of technology ) appearing any time soon, unless someone comes out with a cheap ) scanning laser rangefinder (currently costing around $2000 each) or ) alternative types of sensor are used, such as stereo vision. ) ) ) ) On 10/05/07, J Storrs Hall, PhD [EMAIL PROTECTED] wrote: ) article in NS about the Purdue guessing robot navigators... ) ) http://www.newscientisttech.com/article/dn11805-guessing-robots-navigate-faster.html ) ) I think I get a toljaso on this one --- if the architecture were composed of ) modules that did CBR, each in its own language, from the very start, this ) technique would have fallen out automatically. ) ) 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] Guessing robots
On 10/05/07, Bo Morgan [EMAIL PROTECTED] wrote: You could probably buy 10 cheap webcams and put them all around the robot and get some vision algorithms to turn them into 3D scenes, which are avoided/mapped? This seems like a pretty well understood and constrained problem. This kind of camera based system capable of producing good colour 3D models has just begun to become feasible within the last couple of years. At the moment lasers still remain fashionable amongst researchers since they are more accurate, which means that sensor modelling issues can be ignored for short distance ranging (say up to ten metres), hence the programmer has to do less work. The SLAM algorithms needed to be able to create good quality maps whilst also handling the uncertainty inherent in robot actuators and sensor readings have also been maturing in recent years, to the extent that it's now possible to do reasonable navigation within house sized or laboratory sized areas. - 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] Determinism
David Clark writes: I can predict with high accuracy what I will think on almost any topic. People that can't, either don't know much about the principles they use to think or aren't very rational. I don't use emotion or the current room temperature to make decisions. (No implication that you might ;) Our brains on the microscopic scale might be lossy or non-deterministic but thinking people fix this at the macroscopic level by removing these design defects as quickly as possible from their higher level thinking. Despite rereading this thread I'm not certain what it is even about -- perhaps it is about the difference between a theoretical capability (usually the case when the phrase Turing Machine pops up) and a practical application. However, I don't know if I am the only one, but I do not know with high accuracy what I will think on almost any topic. Just trying to understand what that sentence even means makes my head hurt. I can't predict what I will want for lunch, much less which of the current crop of presidential candidates I prefer (something I have not thought about) or any other nontrivial thing. That may imply that I am irrational; I'd accept that. It feels like squeezing out logical rational inferences is usually difficult and is rarely the way I go about my daily existence. I bother posting this only so you know that not every GI thinks the way you apparently do. - 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
[agi] Open-Source AGI
Just been looking at the vids. of last year's AGI conference. One thing really hit me from the panel talk - and that was: but, of course, only open-source AGI will ever work. Sorry, but all these ideas of individual systems, produced by teams of - what? - say, twenty individuals at most - achieving some significant form of intelligence are, frankly, wild fantasies. We're talking re the human system about the most fabulously complex machine in the universe - and even a simple worm is mind-blowingly complex. Hey, a single cell is awesome. Not just complex as in having many parts but complicated as in having many subsystems. If you stand back and look at AI/ AGI and robotics, as a whole, what you already have anyway is a de facto division of labour of the problem, however crude - different groups are, in fact, going for different aspects of the problem, Emotions, navigation, proprioception, vision, etc. etc. And, you have different roboticists tackling more or less every stage of evolution - with robots from worms and snakes to humanoids. The greatest challenge - and these are my first, very stumbling thoughts here - is to find ways that people can work together on the overall roblem - that all these systems (or subsystems) that people are working on can connect and evolve together. That's the only way that even an adaptive robotic worm [or equivalent] will be produced. (And a common systems/ common parts approach is after all that used by evolution itself). Open-source creativity is the defining model of creativity in this century. The Human Genome Project provided the template not just for biology but for human creativity. And actually, the real singularity - the greatest leap forward - in this century, long before any form of machine intelligence, will be the leap in human creativity that is coming. The last century was that of universal education, this will be the century of universal creativity. Of course, the problem was relatively easy to define for the Human Genome Project. Defining and carving up the problem of AGI so that many teams and the whole world can work on it jointly, is a huge challenge in itself. But it can be done. (Stan Franklin, for example, talked of the problem of just achieving a common ontology, or terminology for AGI, and yet, if you think about it, people ARE using a great deal of common terminology anyway) Ben, I imagine, more or less knows the open-source truth in talking about an AGI Manhattan Project. But even that would be too small. The whole world - the whole Internet - will have to be involved.. - 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] Determinism
--- David Clark [EMAIL PROTECTED] wrote: - Original Message - From: Matt Mahoney [EMAIL PROTECTED] To: agi@v2.listbox.com Sent: Wednesday, May 09, 2007 7:10 PM Subject: Re: [agi] Determinism By simulate, I mean in the formal sense, as a universal Turing machine can simulate any other Turing machine, for example, you can write a program in C that runs programs written in Pascal (e.g. a compiler or interpreter). Thus, you can predict what the Pascal program will do. Languages like Pascal and C define Turing machines. They have unlimited memory. Real machines have finite memory, so you do the simulation properly you need to also define the hardware limits of the target machine. So if the real program reports an out of memory error, the simulation should too, at precisely the same point. Now if the target machine (running Pascal) has 2 MB memory, and your machine (running C) has 1 MB, then you can't do it. Your simulator will run out of memory first. My first computer had 32kb of memory. I know all about doing a lot in very little memory. Lack of memory is about losing time not about the size of physical memory. Virtual memory can be used so that small real memories can simulate much bigger memories. People can simulate absolutely huge systems by just running and looking at small parts of a system at a single time. Your argument might be true IF you talked about full REAL TIME simulation but in general your argument is false. Perhaps I did not state clearly. I assume you are familiar with the concept of a universal Turing machine. Suppose a machine M produces for each input x the output M(x) (or {} if it runs forever). We say a machine U simulates M if for all x, U(m,x) = M(x), where m is a description of M. One may construct universal Turing machines, such that this is true for all M and all x. You can think of U as predicting what M will output for x, without actually running M. In this sense, U can predict its own computations, e.g. U(u,x) = U(x) for all x, where u is a description of U. In other words, U can simulate itself. Turing machines have infinite memory. Real computers have finite memory. There is no such thing as a universal finite state machine. If a machine M has n states (or log2(n) bits of memory), it is not possible to construct a machine U with less than n states such that U(m,x) = M(x) for all x. For some x, yes. I assume that is what you mean. This has nothing to do with speed, and makes no distinction between memory in RAM or on disk. My example used the output from the formula of a line which produces infinite results from only a Y intercept and a slope. You didn't bother to show how my analogy was incorrect. I didn't understand how it was relevant. I can predict with high accuracy what I will think on almost any topic. People that can't, either don't know much about the principles they use to think or aren't very rational. You can't predict when you will next think of something, because then you are thinking of it right now. Maybe you can predict some of your future thoughts, but not all of them. Your brain has finite memory. The best you can do is use a probabilistic approximation of your own thought processes. -- Matt Mahoney, [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=231415user_secret=fabd7936
Re: [agi] Open-Source AGI
Ben, I imagine, more or less knows the open-source truth in talking about an AGI Manhattan Project. But even that would be too small. The whole world - the whole Internet - will have to be involved.. I don't really agree with this. A Manhattan project would be awesome and would maximize odds of success ... but I'm confident that with brilliance, a good AGI design and a bit of luck, a small team can get to the finish line ;-) Ben - 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] Open-Source AGI
On 5/11/07, Mike Tintner [EMAIL PROTECTED] wrote: The greatest challenge - and these are my first, very stumbling thoughts here - is to find ways that people can work together on the overall roblem - that all these systems (or subsystems) that people are working on can connect and evolve together. Agreed. And to do this we need to create a framework/language in which such software is most naturally expressed in a way that's casually reusable - in which apply Alice's genetic algorithm to Bob's vision system on Carol's test set, and distribute the computation across all Dave's spare lab machines whose overnight cycles he's volunteered (where the people involved had not been aware of each other's existence when doing their work) becomes as casual an act as calling sqrt() is today. - 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] Open-Source AGI
On May 10, 2007, at 6:29 PM, Benjamin Goertzel wrote: Ben, I imagine, more or less knows the open-source truth in talking about an AGI Manhattan Project. But even that would be too small. The whole world - the whole Internet - will have to be involved.. I don't really agree with this. A Manhattan project would be awesome and would maximize odds of success ... but I'm confident that with brilliance, a good AGI design and a bit of luck, a small team can get to the finish line ;-) I tend to agree. Many hands and eyeballs are great for a project of many relatively isolatable components whose requirements and interaction are relatively understood. But AGI is pushing the envelope tremendously and, to the degree I understand current designs and design strategies, a set of very tightly inter-related parts need to be designed and build. Many of the parts themselves much less their interaction are being created and integrated out of whole cloth. Small, high bandwidth, concentrated and brilliant teams are required.The vast majority of all programmers/hackers are not qualified. Even of the number that is only a small subset can be formed into a cohesive enough team for this intense a task. If anything is likely to be a natural cathedral rather than a bazaar it is AGI. - samantha - 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] Open-Source AGI
On 5/11/07, Samantha Atkins [EMAIL PROTECTED] wrote: I tend to agree. Many hands and eyeballs are great for a project of many relatively isolatable components whose requirements and interaction are relatively understood. But AGI is pushing the envelope tremendously and, to the degree I understand current designs and design strategies, a set of very tightly inter-related parts need to be designed and build. Many of the parts themselves much less their interaction are being created and integrated out of whole cloth. Small, high bandwidth, concentrated and brilliant teams are required.The vast majority of all programmers/hackers are not qualified. Even of the number that is only a small subset can be formed into a cohesive enough team for this intense a task. If anything is likely to be a natural cathedral rather than a bazaar it is AGI. Well there are two phases, framework and content. The framework is as you say: it needs to be a cathedral. The content needs to be of volume such that only a whole industry can create it: definitely a bazaar. The hard part then is designing a framework such as to allow content to easily flow together. Compare it to the Web: the first browser was created by an individual or small team, but the Web itself was not. - 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] Open-Source AGI
On May 10, 2007, at 6:49 PM, Russell Wallace wrote: Well there are two phases, framework and content. The framework is as you say: it needs to be a cathedral. The content needs to be of volume such that only a whole industry can create it: definitely a bazaar. The hard part then is designing a framework such as to allow content to easily flow together. Compare it to the Web: the first browser was created by an individual or small team, but the Web itself was not. I think (could be wrong) that part of the goal of the core team is to create a mind that can largely navigate huge amounts of data for itself, something that has the basis to learn autonomously on the Web. It may take a phase of a lot of input from many hands to get there but then again it may not. - samantha - 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] Open-Source AGI
On 5/11/07, Samantha Atkins [EMAIL PROTECTED] wrote: I think (could be wrong) that part of the goal of the core team is to create a mind that can largely navigate huge amounts of data for itself, something that has the basis to learn autonomously on the Web. It may take a phase of a lot of input from many hands to get there but then again it may not. *nods* That is certainly an idea, one that I myself spent quite a while studying, and not with a view to disproving it. Alas it doesn't work; the contents of the Web are not information unless you already have a great deal of knowledge beforehand (the old problem of learning Chinese with only a Chinese-Chinese dictionary); and that previously required knowledge is precisely the content that will require an industry, not just a team, to create. - 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] Open-Source AGI
Open source vs closed source is one of the most difficult decisions I faced in my entire AGI career. I've always championed open source AND for-profit, which is the middleground of open-free and closed-commercial, though it may seem like a contradiction. Sometimes I think it may work in a paradoxical way. Some other times, I have a feeling that such a middleground is not the best. One surprising thing I learned is how many AGI people actually insist on the product being free and opensource. Perhaps it is the backlash created by Bill Gates' extraordinary success and wealth, at the expense of his competitors and other startups. * * * The second question is whether the AGI core can be built by a closed team of say 10-20 people. I think the answer is yes and no. It depends on what kind of people we're talking about. AGI requires solving some *open research problems* such as probabilistic logic. Normally any one of such problems requires at least several years of a devoted and brilliant researcher to solve. Notice that this situation is distinctly different from many other traditional startups where basically no major technological breakthrough / research is required (most notably Microsoft). So it seems that we need to INCREASE the level of ideas-sharing and collaboration among researchers. And opensource *may* be able to help achieve that. But then again, notice that most traditional opensource projects are very technologically conservative -- they're not known for bring about breakthroughs. We have tough problems ahead, to say the least. YKY - 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] Open-Source AGI
Mind you, the free/commercial and closed/open-source decisions are separate ones. They're strategic decisions; there's nothing about the problem that intrinsically requires either, it's a matter of coming up with a strategy that can let the participants pay the rent while they work on the project, without compromising its usefulness. That's separate from the technical necessity for focused framework, industry-wide content. - 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] Open-Source AGI
Mike Tintner wrote: The greatest challenge - and these are my first, very stumbling thoughts here - is to find ways that people can work together on the overall problem - that all these systems (or subsystems) that people are working on can connect and evolve together. http://en.wikipedia.org/wiki/Artificial_intelligence_systems_integration says that [I]ntegrating what's already available is a more logical approach to broader A.I. than building monolithic systems from scratch. That's the only way that even an adaptive robotic worm [or equivalent] will be produced. (And a common systems/ common parts approach is after all that used by evolution itself). http://mind.sourceforge.net/aisteps.html breaks the AGO problem down into discrete mind-modles for specialists to work on. ATM -- http://www.advogato.org/person/mentifex/ - 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] Determinism
- Original Message - From: Matt Mahoney [EMAIL PROTECTED] To: agi@v2.listbox.com Sent: Thursday, May 10, 2007 6:04 PM Subject: Re: [agi] Determinism Perhaps I did not state clearly. I assume you are familiar with the concept of a universal Turing machine. Suppose a machine M produces for each input x the output M(x) (or {} if it runs forever). We say a machine U simulates M if for all x, U(m,x) = M(x), where m is a description of M. One may construct universal Turing machines, such that this is true for all M and all x. You can think of U as predicting what M will output for x, without actually running M. In this sense, U can predict its own computations, e.g. U(u,x) = U(x) for all x, where u is a description of U. In other words, U can simulate itself. I could care less about Turing machines or infinite memories. If you want to create an AGI, you will have to use real life computers and real life software, not imaginary musings. Turing machines have infinite memory. Real computers have finite memory. There is no such thing as a universal finite state machine. If a machine M has n states (or log2(n) bits of memory), it is not possible to construct a machine U with less than n states such that U(m,x) = M(x) for all x. For some x, yes. I assume that is what you mean. This has nothing to do with speed, and makes no distinction between memory in RAM or on disk. My example used the output from the formula of a line which produces infinite results from only a Y intercept and a slope. You didn't bother to show how my analogy was incorrect. I didn't understand how it was relevant. A simulation can be as simple as the formula for a line. The formula is the algorithm that defines the simulation and the input uses this formula to produce results. This seems pretty simple to me. With the correct algorithm that has minimal memory or storage requirements, you get an infinite set of answers. This is certainly a class of models. The human brain is defined by a small set of instructions encoded by DNA and this produces the hugely complex brain. Small input, huge output. The memory requirement for a simulation is not proportional to the volume of output. I can predict with high accuracy what I will think on almost any topic. People that can't, either don't know much about the principles they use to think or aren't very rational. You can't predict when you will next think of something, because then you are thinking of it right now. Maybe you can predict some of your future thoughts, but not all of them. Your brain has finite memory. The best you can do is use a probabilistic approximation of your own thought processes. I never said I could PREDICT what I would think at some time in the future, only what I would conclude if I thought about some particular problem. If you told me I said XYZ 5 years ago, I could tell you with absolute accuracy if in fact I did say XYZ or not. The reason is that I am meticulously consistent in the conclusions I draw based on the information I have. This knowledge of what I know and how I think is not probabilistic or approximate. It is totally deterministic and intentional regardless of the inherent non-determinism of my human brain. David Clark - 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