Re: [agi] database access fast enough?
On 4/18/08, J. Andrew Rogers [EMAIL PROTECTED] wrote: On Apr 17, 2008, at 3:32 PM, YKY (Yan King Yin) wrote: Disk access rate is ~10 times faster than ethernet access rate. IMO, if RAM is not enough the next thing to turn to should be the harddisk. Eh? Ethernet latency is sub-millisecond, and in a highly tuned system approaches the 10 microsecond range for something local. Much, much faster than disk if the remote node has your data in RAM and is relatively local. Note that relatively local can mean geographically regional. The round-trip RAM access time from my machine to a machine on the other side of town is a fraction of millisecond over the Internet connection (not hypothetical, actually measured at ~400 microseconds). I wish disk access was even remotely that good. And this was with inexpensive Gigabit Ethernet. LOL... you're right, I forgot to consider latency. Ethernet is much faster than harddisk if we measure access times. But there is another factor: Harddisk is owned by the user. Memory over the net is owned by others, so must be shared. It's not easy to arrange a distributed and cooperative storage scheme. It's hard enough to solve core AGI problems, I simply don't have time to do deal with that. Solid State Disks seems to be a promising solution. YKY --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] database access fast enough?
Plus, learning requires that we store a lot of hypotheses. Let's say 1000-1 times the real KB. I reject this hypothesis as ludicrously incorrect. - Original Message - From: YKY (Yan King Yin) [EMAIL PROTECTED] To: agi@v2.listbox.com Sent: Thursday, April 17, 2008 4:58 PM Subject: Re: [agi] database access fast enough? On 4/18/08, Mark Waser [EMAIL PROTECTED] wrote: Yes. RAM is *HUGE*. Intelligence is *NOT*. Really? I will believe that if I see more evidence... right now I'm skeptical. And your *opinion* has what basis? Are you arguing that RAM isn't huge? That's easily disprovable. Or are you arguing that intelligence is huge? That too is easily disprovable. Which one do I need to knock down? The current OpenCyc KB is ~200 Mbs (correct me if I'm wrong). The RAM size of current high-end PCs is ~10 Gbs. My intuition estimates that the current OpenCyc is only about 10%-40% of a 5 year-old human intelligence. Plus, learning requires that we store a lot of hypotheses. Let's say 1000-1 times the real KB. That comes to 500Gb - 20Tb. It seems that if we allow several years for RAM size to double a few times, RAM may have a chance to catch up to the low end. Obviously not now. YKY --- 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] database access fast enough?
I agree with your side of the debate about whole KB not fitting into RAM. As a solution, I propose to partition the whole KB into the tiniest possible cached chunks, suitable for a single agent running on a host computer with RAM resources of at least one GB. And I propose that AGI will consist not of one program running on one computer, but a vast multitude of separately hosted agents working in concert. Um. Neither side is arguing that the whole KB fit into RAM. I'm arguing that the necessary *core* for intelligence plus enough cached chunks (as you phrase it) to support the current thought processes WILL fit into RAM. It's obviously ludicrous that all the world's knowledge is going to fit into RAM at one time. - Original Message - From: Stephen Reed To: agi@v2.listbox.com Sent: Thursday, April 17, 2008 5:20 PM Subject: Re: [agi] database access fast enough? YKY, I agree with your side of the debate about whole KB not fitting into RAM. As a solution, I propose to partition the whole KB into the tiniest possible cached chunks, suitable for a single agent running on a host computer with RAM resources of at least one GB. And I propose that AGI will consist not of one program running on one computer, but a vast multitude of separately hosted agents working in concert. But my opinion of the OpenCyc concept coverage with respect to that of a human five-year old differs greatly from yours. I concede that 20 OpenCyc facts are about the number a child might know, but in order to properly ground these concepts, I believe that a much larger number of feature vectors will have to be stored or available in abstracted form. For example, there is the concept of the child's mother. Properly grounding that one concept might require abstracting features from thousands of observations: a.. wet hair mother b.. far away mother c.. angy mother d.. mother hidden from view e.. mother in a crowd f.. mother's voice g.. mother in dim light h.. mother from below i.. and so on Of course you can ignore fully grounded concepts as does current Cycorp for its applications, and as I will with Texai until it is past the bootstrap stage. -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 - Original Message From: YKY (Yan King Yin) [EMAIL PROTECTED] To: agi@v2.listbox.com Sent: Thursday, April 17, 2008 3:58:43 PM Subject: Re: [agi] database access fast enough? On 4/18/08, Mark Waser [EMAIL PROTECTED] wrote: Yes. RAM is *HUGE*. Intelligence is *NOT*. Really? I will believe that if I see more evidence... right now I'm skeptical. And your *opinion* has what basis? Are you arguing that RAM isn't huge? That's easily disprovable. Or are you arguing that intelligence is huge? That too is easily disprovable. Which one do I need to knock down? The current OpenCyc KB is ~200 Mbs (correct me if I'm wrong). The RAM size of current high-end PCs is ~10 Gbs. My intuition estimates that the current OpenCyc is only about 10%-40% of a 5 year-old human intelligence. Plus, learning requires that we store a lot of hypotheses. Let's say 1000-1 times the real KB. That comes to 500Gb - 20Tb. It seems that if we allow several years for RAM size to double a few times, RAM may have a chance to catch up to the low end. Obviously not now. YKY --- 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 -- Be a better friend, newshound, and know-it-all with Yahoo! Mobile. Try it now. -- agi | Archives | Modify Your Subscription --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] database access fast enough?
--- Mark Waser [EMAIL PROTECTED] wrote: Um. Neither side is arguing that the whole KB fit into RAM. I'm arguing that the necessary *core* for intelligence plus enough cached chunks (as you phrase it) to support the current thought processes WILL fit into RAM. It's obviously ludicrous that all the world's knowledge is going to fit into RAM at one time. What is your estimate of the quantity of all the world's knowledge? (Or the amount needed to achieve AGI or some specific goal?) Google probably keeps a copy of the searchable part of the internet in about 1 PB of RAM, but this isn't AGI yet. I suppose an internet-wide distributed system could cache about 1 EB (10^18 bytes). -- Matt Mahoney, [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] database access fast enough?
On 4/18/08, Mark Waser [EMAIL PROTECTED] wrote: Um. Neither side is arguing that the whole KB fit into RAM. I'm arguing that the necessary *core* for intelligence plus enough cached chunks (as you phrase it) to support the current thought processes WILL fit into RAM. It's obviously ludicrous that all the world's knowledge is going to fit into RAM at one time. Then we have no disagreement. Notice that the loading-on-demand chunks require that we *duplicate* data. For example facts about JK Rowling can be in a literature chunk as well as a entrepreneur chunk. The question is whether DBMSs support this. Materialized views may be the answer (http://en.wikipedia.org/wiki/Materialized_view). As I said before, minimizing disk access is still an important issue. And all this is peripheral to AGI. I wish I can just focus on AGI algorithms! YKY --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] database access fast enough?
On 4/18/08, Matt Mahoney [EMAIL PROTECTED] wrote: What is your estimate of the quantity of all the world's knowledge? (Or the amount needed to achieve AGI or some specific goal?) Matt, The world's knowledge is irrelevant to the goal of AGI. What we need is to build a commonsense AGI and then let the it control other expert systems with specialized knowledge. So the pertinent question is how large is the core commonsense KB. I guess anywhere from 1Gb to 100Gb is possible, excluding hypotheses from learning, and episodic memory. YKY --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] database access fast enough?
--- Mark Waser [EMAIL PROTECTED] wrote: What is your estimate of the quantity of all the world's knowledge? (Or the amount needed to achieve AGI or some specific goal?) I have no idea (and the question is further muddled by what knowledge is and what formats are included). The question itself is fundamentally nonsensical in it's current form. I mean either algorithmic complexity, or more practically, how much memory you need (which depends on the data representation). But really, it depends on the goal, which I have been trying unsuccessfully for years to get YKY to pin down. -- Matt Mahoney, [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] An Open Letter to AGI Investors
Benjamin Johnston wrote: I have stuck my neck out and written an Open Letter to AGI (Artificial General Intelligence) Investors on my website at http://susaro.com. All part of a campaign to get this field jumpstarted. Next week I am going to put up a road map for my own development project. Hi Richard, If I were a potential investor, I don't think I'd find your letter convincing. AI was first coined some 50 years ago: before I was born, and therefore long before I entered the field of AI. Naturally, I can't speak with personal experience on the matter, but when I read the early literature on AI or when I read about field's pioneers reminiscing on the early days, I get the distinct impression that this was an incredibly passionate and excited group. I would feel comfortable calling them a gang of hot-headed revolutionaries - even today, 50 years after inventing the term AI and at the age of 80, McCarthy writes about AI and the possibility of strong AI with passion and excitement. Yet, in spite of all the hype, excitement and investment that was apparently around during that time (or, more likely, as a result of the hype and excitement) the field crashed in the AI winter of the 80s without finding that dramatic breakthrough. There's the Japanese Fifth Generation Computer Systems project that I understand to be a massive billion dollar investment during the 80s into parallel machines and artificial intelligence; an investment that is today largely considered to be a huge failure. And of course, there's Cyc; formed with an inspiring aim to capture all commonsense knowledge, but still remains in development some 20 years later. And in addition to these, there are the many many early research papers on AI problem solving systems that show early promise and cause the authors to make wild predictions and claims in their Future Work... predictions that time has reliably proven to be false. So, why would I want to invest now? When I track down the biographies of several of the regulars on this list, I find that they entered the field during or after the AI Winter and never experienced the early optimism as an insider. How can you convince an investor that the passion today isn't just the unfounded optimism of researchers who don't remember the past? How can you convince an investor that AGI isn't also going to devolve again into an emphasis on publications rather than quality (as you claim AI has devolved) or into a new kind of weak AGI with no dramatic breakthrough? I think a better argument would be to point to a fundamental technological or methodological change that makes AGI finally credible. I'm not convinced that being lean, mean, hungry and hellbent on getting results is enough. If I believe in AGI, maybe my best bet is to invest my money elsewhere and wait until the fundamental attitudes have changed so each dollar will have a bigger impact, rather than squandered on a bad dead-end idea. Alternately, my best bet may be to invest in weak AI because it will give me a short-term profit (that can be reinvested) AND has a plausible case for eventually developing into strong AI. If you can offer no good reason to invest in AGI today (given all its past failures), aside from a renewed passion of its researchers, then a sane reader would have to conclude that AGI is probably a bad investment. Personally, I'm not sure what I feel about AGI (though, I wouldn't be here if I didn't think it was valuable and promising). However, in this email I'm trying to play the devil's advocate in response to your open letter to investors. Ben, Thanks for the thoughtful comments. I have three responses. First, I think there is a world of difference between passionate researchers at the beginning of the field, in 1956, and passionate researchers in 2008 who have a half-century of other people's mistakes to learn from. The secret of success is to try and fail, then to try again with a fresh outlook. That exactly fits the new AGI crowd. Second, when you say that a better argument would be to point to a fundamental technological or methodological change that makes AGI finally credible I must say that I could not agree more. That is *exactly* what I have tried to do in my project, because I have pointed out a problem with the way that old-style AI has been carried out, and that problem is capable of neatly explaining why the early optimism produced nothing. I have also suggested a solution to that problem, pointed out that the solution has never been tried before (so it has the virtue of not being a failure yet!), and also pointed out that the proposed solution resembles some previous approaches that did have sporadic, spectacular success (the early work in connectionism). However, in my Open Letter post, I did not want to emphasize my own work (I will do that elsewhere on the website), but instead point out some general facts about all AGI
Re: [agi] An Open Letter to AGI Investors
Mark Waser wrote: Richard Loosemore wrote: To say to an investor that AGI would be useful because we could use them to build travel agents and receptionists is to utter something completely incoherent. Not at all. It is catering to their desires and refraining from forcibly educating them. Where is the harm? It's certainly better than getting the door slammed in your face. I think this is a mistake. Selling investors the idea of replacement travel agents and housemaids is something that they know, in their gut, is a stupid idea IN THIS CONTEXT. The context is that you are saying that you will build something with the completely general powers of thought that a person has. If you can build such a thing, then claiming that it will be used for a trivial task after (e.g.) $100 million of development money would make no business sense whatsoever. A big part of being coherent in front of an investor is being able to think your idea through to its logical conclusion. Trying to soft-pedal the idea and pretend that it will be less useful than it really is is considered to be just as bad as overselling the idea - this is thinking too small. Either way, you look as if you haven't really thought it through. Here is what I would call thinking it through. The definition of AGI is that it has all the powers of thought that we have, rather than being able to answer questions about a blocks world perfectly, but be completely incapable of talking about the weather. We all agree on this, no? With that understood, there are some obvious consequences to building an AGI. One is that we will be able to duplicate a machine that has acquired expert-level knowledge in its field. This is a stupendous advance on the situation today, obviously, because it means that if an AGI can be taught to reach expert level in some field, it can be duplicated manyfold and suddenly we have a vast army of people pushing back the frontiers together. Now the question is whether it will be so much harder to produce a housemaid than a medical expert. It is not at all obvious that the housemaid or travel agent will be a step on the road. If we can understand how to make something think, why would our efforts happen to land on the intelligence-point that equates to travel agent? Just because this is the kind of work that a human is forced to do when they cannot get anything better, does not mean that this is a natural level of intellectual capacity. The first AGI could just as easily be a blithering idiot, an idiot-savant, a rocket scientist or an unsurpassable genius. To ask it to be a travel agent is to assume that what you build will have a very particular level of intelligence, and be incapable of improvement, so it would beg the question Why would it only reach that level?. I think, in truth, that this talk of using the first AGIs as travel agents and housemaids is based on a weak analysis of what it would mean to produce an early prototype or a step-on-the-road to full AGI. Because we have in our minds this picture of human beings and the way they develop, some people are automatically assuming that an early AGI would be equivalent to a housemaid. What I am saying here is that this is by no means obvious, at the very least. I think that if we can build such thinking machines, we would surely by that stage have come to understand the dynamics of intellectual development in ways that we have no hope of doing today: we will be able to look inside the developing mind and see what factors enable some thinkers to have trouble getting their thoughts together while others zoom on to great heights of achievement. Given that we will be able to do that, we will have much greater chance of being able to produce something that can continue to develop without hitting a roadblock of some kind. In my opinion, what makes a travel agent a travel agent is not a lack of horsepower, but a complicated interaction of drives and social interactions (as well as some contribution from lack of horsepower). A travel agent, in other words, is more like a genius who got stopped along the way, than a person whose brain simply did not have the right design. Richard Loosemore --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
[agi] The Strange Loop of AGI Funding: now logically proved!
PREMISES: (1) AGI is one of the most complicated problem in the history of science, and therefore requires substantial funding for it to happen. (2) Since all previous attempts failed, investors and funding agencies have enough reason to wait until a recognizable breakthrough to put their money in. (3) Since the people who have the money are usually not AGI researchers (so won't read papers and books), a breakthrough becomes recognizable to them only by impressive demos. (4) If the system is really general-purpose, then if it can give an impressive demo on one problem, it should be able to solve all kinds of problems to roughly the same level. (5) If a system already can solve all kinds of problems, then the research has mostly finished, and won't need funding anymore. CONCLUSION: AGI research will get funding when and only when the funding is no longer needed anymore. Q.E.D. :-( 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
On Fri, Apr 18, 2008 at 1:01 PM, Pei Wang [EMAIL PROTECTED] wrote: PREMISES: (1) AGI is one of the most complicated problem in the history of science, and therefore requires substantial funding for it to happen. Potentially, though, massively distributed, collaborative open-source software development could render your first premise false ... (2) Since all previous attempts failed, investors and funding agencies have enough reason to wait until a recognizable breakthrough to put their money in. (3) Since the people who have the money are usually not AGI researchers (so won't read papers and books), a breakthrough becomes recognizable to them only by impressive demos. (4) If the system is really general-purpose, then if it can give an impressive demo on one problem, it should be able to solve all kinds of problems to roughly the same level. (5) If a system already can solve all kinds of problems, then the research has mostly finished, and won't need funding anymore. CONCLUSION: AGI research will get funding when and only when the funding is no longer needed anymore. Q.E.D. :-( 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/?; Powered by Listbox: http://www.listbox.com -- Ben Goertzel, PhD CEO, Novamente LLC and Biomind LLC Director of Research, SIAI [EMAIL PROTECTED] If men cease to believe that they will one day become gods then they will surely become worms. -- Henry Miller --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
Ben Goertzel wrote: On Fri, Apr 18, 2008 at 1:01 PM, Pei Wang [EMAIL PROTECTED] wrote: PREMISES: (1) AGI is one of the most complicated problem in the history of science, and therefore requires substantial funding for it to happen. Potentially, though, massively distributed, collaborative open-source software development could render your first premise false ... Though it is unlikely to do so, because collaborative open-source projects are best suited to situations in which the fundamental ideas behind the design has been solved. Just having a large gang of programmers on an open-source project does not address Pei's point about AGI being the most complicated problem in the history of science. Pei: what I take you to be saying is that the research problem has an unusually high initial overhead. Richard Loosemore. --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
Potentially, though, massively distributed, collaborative open-source software development could render your first premise false ... Though it is unlikely to do so, because collaborative open-source projects are best suited to situations in which the fundamental ideas behind the design has been solved. I believe I've solved the fundamental issues behind the Novamente/OpenCog design... Time and effort will tell if I'm right ;-) 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
Richard, You are right, though the overhead is not mainly money, but time. Of course I don't really believe in my proof, otherwise I'd say that AGI is impossible. ;-) Among the premises I listed, only (1) is not my personal belief, though I know it is assumed by many people. I believe AGI is basically a theoretical problem, which will be solved by a single person or a small group, with little funding. To make impressive demos, the theoretical result will need to be implemented, where the collaborative open-source projects can help. After that, funding will get in to turn the result into applicable technology. Even so, my previous conclusion still holds --- for the people who want to make the key breakthrough, no funding is available until the breakthrough has been made and (may be after years) recognized. Pei On Fri, Apr 18, 2008 at 1:32 PM, Richard Loosemore [EMAIL PROTECTED] wrote: Ben Goertzel wrote: On Fri, Apr 18, 2008 at 1:01 PM, Pei Wang [EMAIL PROTECTED] wrote: PREMISES: (1) AGI is one of the most complicated problem in the history of science, and therefore requires substantial funding for it to happen. Potentially, though, massively distributed, collaborative open-source software development could render your first premise false ... Though it is unlikely to do so, because collaborative open-source projects are best suited to situations in which the fundamental ideas behind the design has been solved. Just having a large gang of programmers on an open-source project does not address Pei's point about AGI being the most complicated problem in the history of science. Pei: what I take you to be saying is that the research problem has an unusually high initial overhead. Richard Loosemore. --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?; 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
Pei: I believe AGI is basically a theoretical problem, which will be solved by a single person or a small group, with little funding How do you define that problem? --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
See http://nars.wang.googlepages.com/wang.AI_Definitions.pdf On Fri, Apr 18, 2008 at 2:11 PM, Mike Tintner [EMAIL PROTECTED] wrote: Pei: I believe AGI is basically a theoretical problem, which will be solved by a single person or a small group, with little funding How do you define that problem? --- 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
On 4/19/08, Richard Loosemore [EMAIL PROTECTED] wrote: PREMISES: (1) AGI is one of the most complicated problem in the history of science, and therefore requires substantial funding for it to happen. Potentially, though, massively distributed, collaborative open-source software development could render your first premise false ... Though it is unlikely to do so, because collaborative open-source projects are best suited to situations in which the fundamental ideas behind the design has been solved. I agree. Opensource is a good thing but it is not sufficient to solve fundamental problems such as architecture and algorithm design. Very few people have comprehensive understanding of AGI, and the few who do are not collaborating, due to theoretical differences. YKY --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
On 4/19/08, Ben Goertzel [EMAIL PROTECTED] wrote:
Though it is unlikely to do so, because collaborative open-source
projects are best suited to situations in which the fundamental ideas behind
the design has been solved.
I believe I've solved the fundamental issues behind the Novamente/OpenCog
design...
It's hard to tell whether you have really solved the AGI problem, at
this stage. ;)
Also, your AGI framework has a lot of non-standard, home-brew stuff
(especially the knowledge representation and logic). I bet there are
some merits in your system, but is it really so compelling that
everybody has to learn it and do it that way?
Creating a standard / common framework is not easy. Right now I think
we lack such a consensus. So the theorists are not working together.
YKY
---
agi
Archives: http://www.listbox.com/member/archive/303/=now
RSS Feed: http://www.listbox.com/member/archive/rss/303/
Modify Your Subscription:
http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4
Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
Another problem is how to judge the impressiveness of a demo, especially if you're a non expert. It's relatively easy to come up with superficially impressive demos, which then turn out upon closer investigation to be fraught with problems or just not scalable. This seems to happen all the time with robotics and computer vision, as countless humanoids show. So I think you're right. The big money only arrives once most of the research problems have been hammered out and a working prototype is available for inspection. Funding might appear earlier only if the organisations involved are suitably convinced that (a) the promised technology is going to arrive in the near future and (b) that there is a strong first mover advantage to owning or influencing that technology. It may, as Ben says, be possible to ameliorate some of the costs using open source methods. Open source is not a panacea, but it could help to turn AGI into more of a science than an art form in that it permits experiments to be independently verified with greater ease and provides more opportunities to stand on the shoulders of giants. On 18/04/2008, Pei Wang [EMAIL PROTECTED] wrote: PREMISES: (1) AGI is one of the most complicated problem in the history of science, and therefore requires substantial funding for it to happen. (2) Since all previous attempts failed, investors and funding agencies have enough reason to wait until a recognizable breakthrough to put their money in. (3) Since the people who have the money are usually not AGI researchers (so won't read papers and books), a breakthrough becomes recognizable to them only by impressive demos. (4) If the system is really general-purpose, then if it can give an impressive demo on one problem, it should be able to solve all kinds of problems to roughly the same level. (5) If a system already can solve all kinds of problems, then the research has mostly finished, and won't need funding anymore. CONCLUSION: AGI research will get funding when and only when the funding is no longer needed anymore. Q.E.D. :-( 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/?; 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
On 4/19/08, YKY (Yan King Yin) [EMAIL PROTECTED] wrote: we lack such a consensus. So the theorists are not working together. I correct that. Theorists do not need to work together; theories can be applied anywhere. It's the *designers* who are not working together. YKY --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
On 18/04/2008, Pei Wang [EMAIL PROTECTED] wrote: I believe AGI is basically a theoretical problem, which will be solved by a single person or a small group, with little funding. I'm not sure I believe this. After working on this a bit, it has become clear to me that there are more ideas than there is time to explore them all. Exploration is further hindered by a lack of software infrastructure. There are no lab facilities, no easy way to perform high-level experiments. I know certainly that I have some high-level theoreies I want ot explore, but I can't even get started due to the lack of infrastructure. I think what Ben is trying to do is to provide those facilities by providing OpenCog. I think opeen-source programmers *can* help build this. And, judging by the Google summer-of-code applications, many of the students have a strong understanding of many of the basic concepts. Richard wrote: Though it is unlikely to do so, because collaborative open-source projects are best suited to situations in which the fundamental ideas behind the design has been solved. Just having a large gang of programmers on an open-source project does not address Pei's point about AGI being the most complicated problem in the history of science. Yes, but a large gang of open source programmers can help build the infrastructure. Curing the Manhattan project, it may have been Feynmann and von Neumann and Teller and Oppenheimer doing all the thinking, but it sure wasn't them that built 42 acres of uranium enrichment plants. This was done by large gangs. The fundamental ideas behind Bayesian nets and whatever have been solved but there is no way, not without a lot of work, to hook Bayesian nets to english language parsers, or to any sort of predicate reasoning systems, or knowledge representation systems or ontologies. Doing such a hookup is scientifically straight-foward and scientifically easy but a huge pain-in-the-arse. Until this hookup is done, we can't run experiments,. can't test theories, can't even get started on solving the scientifically hard part of the problem. --linas --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
Linas Vepstas wrote: Richard wrote: Though it is unlikely to do so, because collaborative open-source projects are best suited to situations in which the fundamental ideas behind the design has been solved. Just having a large gang of programmers on an open-source project does not address Pei's point about AGI being the most complicated problem in the history of science. Yes, but a large gang of open source programmers can help build the infrastructure. Curing the Manhattan project, it may have been Feynmann and von Neumann and Teller and Oppenheimer doing all the thinking, but it sure wasn't them that built 42 acres of uranium enrichment plants. This was done by large gangs. I guess I agree with this point by itself (I could do with a large gang, for example, to build SAFAIRE) but when I made the remarks I was thinking about solving the actual core problem of designing the right architecture. So for example, I think Pei is correct to point out that the basic solution is going to come from one person's idea, but that if we have a situation in which nobody has yet had that idea, then (and only then) the strategy of getting a large gang together would not help. Now, Ben thinks that he does have the correct solution and is ready for the gang. I think the same about my solution, and perhaps Pei Wang and Peter Voss and Hugo de Garis all have the same opinion about their own work in other words, perhaps they all believe that all they need right now is a large enough gang. But if we were all wrong, then our gangs would only serve to prove (eventually!) that we were wrong. I doubt that the open-source collective would be where the solution would come from. So, no disagreement that a big gang is beneficial, but Richard Loosemore P.S. Now I am deep trouble because I just said that a big open source collective could help me build SAFAIRE, and Stephen Reed is going to ask me any minute now why I don't simply get me a big open-source gang ;-) --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
Linas, Not all theoretical problems can or need to be solved by practical testing. Also, in this field, no infrastructure is really theoretically neutral --- OpenCog is clearly not suitable to test all kinds of AGI theories, though I like the project, and is willing to help. Open-source will solve many technical problems, and may also reveal many theoretical problems by putting theories under testing. However, it won't replace theoretical thinking. Pei On Fri, Apr 18, 2008 at 3:59 PM, Linas Vepstas [EMAIL PROTECTED] wrote: On 18/04/2008, Pei Wang [EMAIL PROTECTED] wrote: I believe AGI is basically a theoretical problem, which will be solved by a single person or a small group, with little funding. I'm not sure I believe this. After working on this a bit, it has become clear to me that there are more ideas than there is time to explore them all. Exploration is further hindered by a lack of software infrastructure. There are no lab facilities, no easy way to perform high-level experiments. I know certainly that I have some high-level theoreies I want ot explore, but I can't even get started due to the lack of infrastructure. I think what Ben is trying to do is to provide those facilities by providing OpenCog. I think opeen-source programmers *can* help build this. And, judging by the Google summer-of-code applications, many of the students have a strong understanding of many of the basic concepts. Richard wrote: Though it is unlikely to do so, because collaborative open-source projects are best suited to situations in which the fundamental ideas behind the design has been solved. Just having a large gang of programmers on an open-source project does not address Pei's point about AGI being the most complicated problem in the history of science. Yes, but a large gang of open source programmers can help build the infrastructure. Curing the Manhattan project, it may have been Feynmann and von Neumann and Teller and Oppenheimer doing all the thinking, but it sure wasn't them that built 42 acres of uranium enrichment plants. This was done by large gangs. The fundamental ideas behind Bayesian nets and whatever have been solved but there is no way, not without a lot of work, to hook Bayesian nets to english language parsers, or to any sort of predicate reasoning systems, or knowledge representation systems or ontologies. Doing such a hookup is scientifically straight-foward and scientifically easy but a huge pain-in-the-arse. Until this hookup is done, we can't run experiments,. can't test theories, can't even get started on solving the scientifically hard part of the problem. --linas --- 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
On 4/19/08, Pei Wang [EMAIL PROTECTED] wrote: Not all theoretical problems can or need to be solved by practical testing. Also, in this field, no infrastructure is really theoretically neutral --- OpenCog is clearly not suitable to test all kinds of AGI theories, though I like the project, and is willing to help. Open-source will solve many technical problems, and may also reveal many theoretical problems by putting theories under testing. However, it won't replace theoretical thinking. I agree, but I'd add that it is still tremendously helpful to have an opensource gang solving technical problems. For this reason, I'm tempted to opensource my stuff, but where would be my compensation? Do I really HAVE to sacrifice my pay check...?? YKY --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
Richard, Though I do believe I have the right idea, I surely know that there are still issues I haven't fully solved. Therefore I don't really want a big gang at now (that will only waste the time of mine and the others), but a small-but-good gang, plus more time for myself --- which means less group debates, I guess. ;-) Pei On Fri, Apr 18, 2008 at 4:31 PM, Richard Loosemore [EMAIL PROTECTED] wrote: Now, Ben thinks that he does have the correct solution and is ready for the gang. I think the same about my solution, and perhaps Pei Wang and Peter Voss and Hugo de Garis all have the same opinion about their own work in other words, perhaps they all believe that all they need right now is a large enough gang. But if we were all wrong, then our gangs would only serve to prove (eventually!) that we were wrong. I doubt that the open-source collective would be where the solution would come from. So, no disagreement that a big gang is beneficial, but --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
On Sat, Apr 19, 2008 at 12:48 AM, YKY (Yan King Yin) [EMAIL PROTECTED] wrote: For this reason, I'm tempted to opensource my stuff, but where would be my compensation? Do I really HAVE to sacrifice my pay check...?? Yes, you do, as Wang's Theorem demonstrates. You must persevere in your Faith, and the Way to Nerds' Heaven will open to you, after years of poverty-stricken life as underfunded AGI researcher. :-) -- Vladimir Nesov [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
Pei: I don't really want a big gang at now (that will only waste the time of mine and the others), but a small-but-good gang, plus more time for myself --- which means less group debates, I guess. ;-) Alternatively, you could open your problems for group discussion think-tanking... I'm surprised that none of you systembuilders do 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
On Fri, Apr 18, 2008 at 5:35 PM, Mike Tintner [EMAIL PROTECTED] wrote: Pei: I don't really want a big gang at now (that will only waste the time of mine and the others), but a small-but-good gang, plus more time for myself --- which means less group debates, I guess. ;-) Alternatively, you could open your problems for group discussion think-tanking... I'm surprised that none of you systembuilders do this. That is essentially what I'm doing with OpenCog ... but it's a big job, just preparing stuff in terms of documentation and code and designs so that others have a prayer of understanding it ... 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=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
YKY, I believe I've solved the fundamental issues behind the Novamente/OpenCog design... It's hard to tell whether you have really solved the AGI problem, at this stage. ;) Understood... Also, your AGI framework has a lot of non-standard, home-brew stuff (especially the knowledge representation and logic). I bet there are some merits in your system, but is it really so compelling that everybody has to learn it and do it that way? I don't claim that the Novamente/OpenCog design is the **only** way ... but I do note that the different parts are carefully designed to interoperate together in subtle ways, so replacing any one component w/ some standard system won't work. For instance, replacing PLN with some more popular but more limited probabilistic logic framework, would break a lot of other stuff... Creating a standard / common framework is not easy. Right now I think we lack such a consensus. So the theorists are not working together. One thing that stuck out at the 2006 AGI Workshop and AGI-08 conference, was the commonality between several different approaches, for instance -- my Novamente approach -- Nick Cassimatis's Polyscheme system -- Stan Franklin's LIDA approach -- Sam Adams (IBM) Joshua Blue -- Alexei Samsonovich's BICA architecture Not that these are all the same design ... there are very real differences ... but there are also a lot of deep parallels. Novamente seems to be more fully fleshed out than these overall, but each of these guys has thought through specific aspects more deeply than I have. Also, John Laird (SOAR creator) is moving SOAR in a direction that's a lot closer to the Goertzel/Cassimatis/Franklin/Adams style system than his prior approaches ... All the above approaches are -- integrative, involving multiple separate components tightly bound together in a high-level cognitive architecture -- reliant to some extent on formal inference (along with subsymbolic methods) -- clearly testable/developable in a virtual worlds setting I would bet that with appropriate incentives all of the above researchers could be persuaded to collaborate on a common AI project -- without it degenerating into some kind of useless committee-think... Let's call these approaches LIVE, for short -- Logic-incorporating, Integrative, Virtually Embodied On the other hand, when you look at -- Pei Wang's approach, which is interesting but is fundamentally committed to a particular form of uncertain logic that no other AGI approach accepts -- Selmer Bringsjord's approach, which is founded on the notion that standard predicate logic alone is The Answer -- Hugo de Garis's approach which is based on brain emulation you're looking at interesting approaches that are not really compatible with the LIVE approach ... I'd say, you could not viably bring these guys into a collaborative AI project based on the LIVE approach... So, I do think more collaboration and co-thinking could occur than currently does ... but also that there are limits due to fundamentally different understandings OpenCog is general enough to support any approach falling within the LIVE category, and a number of other sorts of approaches as well (e.g. a variety of neural net based architectures). But it is not **completely** general and doesn't aim to me ... IMO, a completely general AGI development framework is just basically, say, C++ and Linux ;-) -- Ben G --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
FW: [agi] The Strange Loop of AGI Funding: now logically proved!
In the below quote in the below article the number 1000 was meant to be 100 in the below quote. With intelligent RAM, this number could be perhaps a high as 500, depending on what you mean by a current PC, but intelligent RAM would, at least initially, be much more expensive. Such a system would crank roughly 1TOpp/sec and enable 4G random accesses in its 1TB of DRAM/sec. That allows a fair amount of connectionism to be computed --- more than 1000 times as much as current PCs -Original Message- From: Ed Porter [mailto:[EMAIL PROTECTED] Sent: Friday, April 18, 2008 6:36 PM To: agi@v2.listbox.com Subject: RE: [agi] The Strange Loop of AGI Funding: now logically proved! PEI'S SELF-DEFEATING LOOP WILL PROBABLY BE BROKEN WITHIN 3 TO 8 YEARS Over then next 3 - 8 years there probably will arise from the level of AI and AGI projects being funded an ever increasing appreciation and proof of the power, generality, and potential of AGI --- enough so that funding of large AGI projects will start. This will be aided by the growing increase in the number of people who believe in AGI --- the increasing organization of the AGI community, as represented by the recent AGI 2008 conference and the planed AGI 2009 --- the growing knowledge and tools, such as OpenCog tools, for building AGI projects --- and, perhaps most importantly, the rapidly dropping cost and rapidly rising power of hardware. Today a PC with 4G or RAM should be able to demonstrate some important pieces of the AGI problem. For under $40k --- money many funded projects can afford --- you can buy a 4 processor quad-core server with 256GB DRAM and many TBs of disk. With that you should be able to demonstrate even more of AGI's potential. But in about 5 years things should really start changing with the arrival of the many-core, many-layer chips and the operating systems for efficiently using them (with the help of people like Sam Adams). We should begin to see chips with 256 cores --- connected by a high-speed on chip mesh network --- with each core having fat thru-wafer paths to memory totaling --- for the whole chip --- roughly 8GBytes of DRAM --- plus a GByte of embedded DRAM for L2 cache. All this will fit in one multi-layer chip. Each such chip will have --- say --- 1TByte of external bandwidth in the form of 128 different 8GByte/sec channels to memory or other processors. I'm guessing such chip will probably sell for a very fat premium at introduction --- say $5K --- but will be under $2K within two years. In 5 years a TByte of DRAM should cost about $10K. This means by 2013 you could have a system with such a chip and 1TB of DRAM for under $20K --- cheap enough for most teams of AI grad students to share one. Such a system would crank roughly 1TOpp/sec and enable 4G random accesses in its 1TB of DRAM/sec. That allows a fair amount of connectionism to be computed --- more than 1000 times as much as current PCs. Such a $20K system should be a powerful enough testbed to develop and test much of the basic architecture for artificial minds. And they should be cheap enough that within the next six or seven years hundreds of AGI teams could test and perfect their approaches on them. Such testbeds could come in a range of different sizes --- including one with 256 such 256 core chips and 50TBytes of DRAM. This would include 2TB of on-chip DRAM and 256G of L2 cache. Such hardware could provide 100 TOpps --- 512G random accesses/sec in the 50TB of DRAM --- and a theoretical cross-sectional bandwidth between chips of about 512G global 64byte messages/sec (with a much higher total number of interchip messages/sec, since a many messages would be between near by chips). This computational, representational, and communication capacity is very possibly enough to create human level AGI. In 7 years such hardware could cost roughly $1m in parts, which means it probably could be profitable be sold for under $2m if it was being sold in any quantity. And understand these system would be good for general scientific, data base, virtual reality, and advanced web-based programming as well --- so they should sell in quantity --- and their price should come down rapidly. At these price every major university and research lab could afford several of them. Within 10 years from today there could be thousands such roughly human level machines. The combination of increasingly interest in AGI, increasing understand of it, increasingly powerful AGI tools --- and more powerful hardware --- all will combine to make it highly likely that within 3 to 8 years there will have been enough progress that it will become obvious that there is tremendous strategic and economic value for investing in it big time --- and then the big money will come in the hundreds of millions or billions of dollars. So I think Pei's self defeating strange loop of AGI funding holds in the short term --- but also that in less than a decade this self-defeating loop will
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
--- Pei Wang [EMAIL PROTECTED] wrote: I believe AGI is basically a theoretical problem, which will be solved by a single person or a small group, with little funding. I think that we are still massively underestimating the cost of AGI, just as we have been doing for the last 50 years. The value of AGI is the value of the human labor it would replace, between $2 and $5 quadrillion over the next 30 years worldwide. To suggest that it could be solved for a billionth of this cost is ludicrous. Google has $169 billion and the motivation, market, brains, and computing power to solve AGI, but they haven't yet. I realize there is a tradeoff between having AGI sooner or waiting for the price of technology to come down. Simple economics suggest we will be willing to pay a significant fraction of the value to have it now. We are chasing a moving target. It is not enough for a computer to match the intelligence of a human. It has to match the intelligence of a human with an internet connection, and the internet keeps getting smarter as AI is deployed on it. You hit the target not at one human brain (which Google has probably surpassed in computing power and data), but at 10 billion human brains. You need a vision system for a billion eyes, a language model to converse with a billion people at the same time. Given a good communication infrastructure, general models of intelligence are at a distinct disadvantage against narrow AI. You will be competing with millions or billions of specialized experts that are individually easier to build, train, optimize, and maintain for one particular task and run on a PC using mature technology. Standalone AGI can't do that. We don't even know how much computing power is needed to do what one brain does (but I am pretty sure it is more than 1000 PCs). I know the argument that you only have to build it once. I've heard it before. Each standalone AGI has to be trained for a different task. This is a nontrivial expense. We should not expect it to cost significantly less than training a new employee. I think AGI is too big for anyone to own or invest in. If you want to invest, look for market opportunities that don't yet exist, the way Google built its fortune by indexing something that didn't exist 15 years ago. -- Matt Mahoney, [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Re: [agi] The Strange Loop of AGI Funding: now logically proved!
On 4/19/08, Ben Goertzel [EMAIL PROTECTED] wrote: I don't claim that the Novamente/OpenCog design is the **only** way ... but I do note that the different parts are carefully designed to interoperate together in subtle ways, so replacing any one component w/ some standard system won't work. This problem may be common to all AGI designs -- no one seems to be able to build AGI out of standard components. But I would strive towards that ideal as close as possible. For instance, replacing PLN with some more popular but more limited probabilistic logic framework, would break a lot of other stuff... PLN is not based on predicate logic but on term logic, right? That may be a source of problem. I would bet that with appropriate incentives all of the above researchers could be persuaded to collaborate on a common AI project -- without it degenerating into some kind of useless committee-think... THAT would be highly desirable, but are we ready yet to reconcile our differences? I guess we can start from gradually re-using standard components in a bottom-up manner. Also, establishing a knowledge interchange format. Let's call these approaches LIVE, for short -- Logic-incorporating, Integrative, Virtually Embodied LIVE is good =) -- Pei Wang's approach, which is interesting but is fundamentally committed to a particular form of uncertain logic that no other AGI approach accepts I think Pei Wang makes his own versions of abduction and induction that, from the classical logic perspective, are unsound. Otherwise his approach is also LIVE. -- Selmer Bringsjord's approach, which is founded on the notion that standard predicate logic alone is The Answer Agreed. Binary logic can go a long way, but ultimately is insufficient for AGI. That said, I'm currently designing learning algorithms using binary logic only, and plan to add fuzzy and probability later. OpenCog is general enough to support any approach falling within the LIVE category, and a number of other sorts of approaches as well (e.g. a variety of neural net based architectures). But it is not **completely** general and doesn't aim to me ... IMO, a completely general AGI development framework is just basically, say, C++ and Linux ;-) Yes, OpenCog is definitely a good move. I hope you will allow it more free so it can bring about more fundamental changes, to the point that different AGI projects can interoperate. =) YKY --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
Open source (was Re: [agi] The Strange Loop of AGI Funding: now logically proved!)
--- YKY (Yan King Yin) [EMAIL PROTECTED] wrote: On 4/19/08, Pei Wang [EMAIL PROTECTED] wrote: Not all theoretical problems can or need to be solved by practical testing. Also, in this field, no infrastructure is really theoretically neutral --- OpenCog is clearly not suitable to test all kinds of AGI theories, though I like the project, and is willing to help. Open-source will solve many technical problems, and may also reveal many theoretical problems by putting theories under testing. However, it won't replace theoretical thinking. I agree, but I'd add that it is still tremendously helpful to have an opensource gang solving technical problems. For this reason, I'm tempted to opensource my stuff, but where would be my compensation? Do I really HAVE to sacrifice my pay check...?? Not at all. I released most of my data compression software under GPL. If a company wants to use it in a commercial product or wants something customized, they have to pay me. Meanwhile a lot of people have improved the software for free until it moved to the top of the rankings where it got the attention of companies that need data compression experts and pay well. This would never have happened if I had kept the code proprietary. -- Matt Mahoney, [EMAIL PROTECTED] --- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244id_secret=101455710-f059c4 Powered by Listbox: http://www.listbox.com
