Re: [computer-go] Big board, ++physics
Le mercredi 28 février 2007 16:49, Oliver Lewis a écrit : On 2/23/07, David Doshay [EMAIL PROTECTED] wrote: On 22, Feb 2007, at 9:03 PM, alain Baeckeroot wrote: ... I made very slow progress to formalize this ... But the whole stuff is rather coherent in my mind. Then I envy you. I have been trying to bring what I know about MC in physics together with Go for over 20 years, and I get tripped up every time by temperature. I know how to deal with it properly in the physics, but I still have no idea at all about how to cool the MC Go simulations. The concept of temperature as used in CGT (combinatorial game theory) has not helped me. David - using Alain's analogy about temperature being related to mixing, isn't there a link with what Peter Drake calls the proximity heuristic in the MC playouts? A completely random MC player may be too hot and one that always plays next to already occupied points too cold. In between, it should be possible to define a temperature parameter which controls how close to the existing points a random MC playout happens. You could then test how strength varies with this temperature parameter. Is this what either of you had in mind? Yes :) Beginners do not mix enought, random players mix too much. Alain ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
I do agree with Alain that beginners mix too little and random players too much. I am most intrigued with the recent results from Dave Hillis, where he shows what I have been calling a move towards a transition temperature with a selected set of heuristics in the playout. When he is willing to go public with the details of the correspondence between the heuristics and the scaling behavior we will all know more. In the mean time, I am completely hobbled by my physics background and knowledge. I am too stuck with thinking about some kind of energy function to make progress the way Dave is. But I am working towards putting aside any formal sense of what I think is correct and just trying and testing like Dave is ... a sort of random experimental view where I just try stuff and see what happens. Perhaps that will eventually lead to a new insight. Cheers, David On 28, Feb 2007, at 10:02 AM, alain Baeckeroot wrote: Le mercredi 28 février 2007 16:49, Oliver Lewis a écrit : On 2/23/07, David Doshay [EMAIL PROTECTED] wrote: On 22, Feb 2007, at 9:03 PM, alain Baeckeroot wrote: ... I made very slow progress to formalize this ... But the whole stuff is rather coherent in my mind. Then I envy you. I have been trying to bring what I know about MC in physics together with Go for over 20 years, and I get tripped up every time by temperature. I know how to deal with it properly in the physics, but I still have no idea at all about how to cool the MC Go simulations. The concept of temperature as used in CGT (combinatorial game theory) has not helped me. David - using Alain's analogy about temperature being related to mixing, isn't there a link with what Peter Drake calls the proximity heuristic in the MC playouts? A completely random MC player may be too hot and one that always plays next to already occupied points too cold. In between, it should be possible to define a temperature parameter which controls how close to the existing points a random MC playout happens. You could then test how strength varies with this temperature parameter. Is this what either of you had in mind? Yes :) Beginners do not mix enought, random players mix too much. Alain ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
One more thought: It would be interesting to see the degree to which following a proximity heuristic leads to the renormalizations looking cold. Cheers, David On 28, Feb 2007, at 11:07 AM, David Doshay wrote: I do agree with Alain that beginners mix too little and random players too much. I am most intrigued with the recent results from Dave Hillis, where he shows what I have been calling a move towards a transition temperature with a selected set of heuristics in the playout. When he is willing to go public with the details of the correspondence between the heuristics and the scaling behavior we will all know more. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
On Tue, 2007-02-27 at 11:11 +0900, igo wrote: If computers ever become world champion strength at 19x19, there will probably have been some simplification that makes this possbile, I don't see it being a (direct) result of faster computers or more processors. So in this situation it is POSSIBLE, that the game gets difficult more quickly for humans than for computers if some such wonderful discovery is made. It is an impressive idea, thanks. Sometimes I also imagine that a computer connect to a brain. Then the computer would be a so-called 'fifty-fifty' machine^^. Then it would be a fast way to overcome human easily anytime, any boardsize. I've often imagined a computer brain implant that would allow you to seamlessly use the computing facilities of a computer, but with a very natural interface, input and output directly to the brain. igo ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
Hello Jacque: Thanks for the comments. my point is that 19x19 is the optimal size for human abilities. I don't think so. 19x19 is merely the size of Go originally. for human abilities in Go, 19x19, 21x21...99x99 are about the same. ... The entire fuseki theory is board size dependent. Not really. fuseki theory is logic dependent. When human learns Go, he learns the logic of Go, the essence of Go, but doesn't care about the boardsize so much. (I am human, and you are human too. I think^^) Sometimes when boardsize is involved in some situation, then human will adjust it automatically, like adjust temperature in winter or summer. Top humans playing 21x21 may be weaker than a 19x19 world champion computer. May be, but it will take time (to proof that) surely. My point is simple. for example, [MoGo] can beat a 3d person at 9x9 now. but the same person(3d) will beat [MoGo] at 13x13 easily at this time. Will you agree ? when [MoGo] can beat the same person at 13x13, then the same person will beat [MoGo] at 19x19 easily at that time. Will you agree ? The person doesn't improve in Go at all, but only adjusts automatically. igo ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
RE : Re: [computer-go] Big board, ++physics
no englich me french igo [EMAIL PROTECTED] a écrit : Hello Jacque: Thanks for the comments. my point is that 19x19 is the optimal size for human abilities. I don't think so. 19x19 is merely the size of Go originally. for human abilities in Go, 19x19, 21x21...99x99 are about the same. ... The entire fuseki theory is board size dependent. Not really. fuseki theory is logic dependent. When human learns Go, he learns the logic of Go, the essence of Go, but doesn't care about the boardsize so much. (I am human, and you are human too. I think^^) Sometimes when boardsize is involved in some situation, then human will adjust it automatically, like adjust temperature in winter or summer. Top humans playing 21x21 may be weaker than a 19x19 world champion computer. May be, but it will take time (to proof that) surely. My point is simple. for example, [MoGo] can beat a 3d person at 9x9 now. but the same person(3d) will beat [MoGo] at 13x13 easily at this time. Will you agree ? when [MoGo] can beat the same person at 13x13, then the same person will beat [MoGo] at 19x19 easily at that time. Will you agree ? The person doesn't improve in Go at all, but only adjusts automatically. igo ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ - Découvrez une nouvelle façon d'obtenir des réponses à toutes vos questions ! Profitez des connaissances, des opinions et des expériences des internautes sur Yahoo! Questions/Réponses.___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
On Mon, 2007-02-26 at 02:50 +0900, igo wrote: My point is simple. for example, [MoGo] can beat a 3d person at 9x9 now. but the same person(3d) will beat [MoGo] at 13x13 easily at this time. Will you agree ? when [MoGo] can beat the same person at 13x13, then the same person will beat [MoGo] at 19x19 easily at that time. Will you agree ? The person doesn't improve in Go at all, but only adjusts automatically. igo Your point follows the logical progression, so far the bigger the board the better humans are relative to computers. If computers ever become world champion strength at 19x19, there will probably have been some simplification that makes this possbile, I don't see it being a (direct) result of faster computers or more processors. So in this situation it is POSSIBLE, that the game gets difficult more quickly for humans than for computers if some such wonderful discovery is made. This is all pretty speculative though, probably you are right. - Don ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
Hello igo: igo wrote (on behalf of Making the board bigger would probably make the game weaker for humans. I presume the day a computer is world champion, increasing board size would give the computer even more advantage.): I presume exact the opposite way. Of course, who knows. This is only intuition, but my point is that 19x19 is the optimal size for human abilities. It is both complicated and the game can be analyzed very deeply. It has taken centuries for humans to reach near-perfection (another intuition). If computers reach humans after, say 40 years of research, all gigantic-search-space related problems will have been solved. (They are present at much smaller sizes.) Making the search space 22% bigger won't be a big problem compared with creating 21x21 human Go theory from scratch. According to Ikuro Ishigure, professional 8 dan, author of In the Beginning, in a professional game of two days, the first day is dedicated to the first 50 moves. The entire fuseki theory is board size dependent. Fuseki is based on presumptions and those are founded on experience. Top humans playing 21x21 may be weaker than a 19x19 world champion computer. The computer will be probably the champion (as in chess) because it can search deeper and do error free subdivision of the search space. The same groups it can overhumanly kill/defend on a 19x19 board may be managed on a bigger board in similar time. (If it looses its time considering unrelated moves, it won't be a 19x19 world champion either.) Again, who knows. I hope to see that in my lifetime. Jacques. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
Ray Tayek wrote: it's also hard to see why 21x21 would be boring (i can see 17x17 being too simple in some sense). There is also the length of a game. 21x21 is 22% bigger in terms of cells. Professional players can work two days on a 19x19 game. Making the board bigger would probably make the game weaker for humans. I presume the day a computer is world champion, increasing board size would give the computer even more advantage. (Against the common search-width based intuition.) Jacques. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
Making the board bigger would probably make the game weaker for humans. I presume the day a computer is world champion, increasing board size would give the computer even more advantage. (Againstthe common search-width based intuition.) I presume exact the opposite way. The day a computer is world champion at 19x19, increasing boardsize would give human advantage. because computer has mechanical power, but human has wisdom. igo ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
-Original Message- From: [EMAIL PROTECTED] To: computer-go@computer-go.org Sent: Fri, 23 Feb 2007 12:03 AM Subject: Re: [computer-go] Big board, ++physics Your analogy with physics encourage me to share other physical analogies. 1/ Cooling the simulation could be done by controlling the mixing rate and the density of stones. -Beginners'games are too cold, not enought mixed (=overconcentrated or very high viscosity, nearly solid state, not ignitable) -Professionnal games are probably near critical state (explosive conditions, gaz state) -MC-players are nearly random = too hot, too mixed, plasma state. 2/ Soap Bubbles = potential territory In addition to previous fluid state, i see hypothetical bubbles: - beginners makes some (less than 10) big bubbles, and their size and place are early known. (still too cold and too high viscosity) - professional can makes lots of bubbles (20+), but they are changing and turning very often and quickly - nearly-random makes a foam 3/ Solidification and cristal growth often comes to mind. Cristal growth need a seed to begin, generally it is a defect or some impurity. In go the defect are the corners: - they need less material to build a frontier (like soap bubbles) so corners are the beginning of the process of solidification or cristal growth. - the topology of the corner (2 libs, 3 libs and 4 libs) imposes the size and shape of a living group. - impurity is a captured stone/group 4/ shape/size resonance (un)fortunately the 19x19 size is just the critical size to have problems. -17x17 is too small, corners influence is too strong, it is quickly possible to take the border. (= 3 bubbles) -21x21 is too wide, it is not possible to quickly prevent easy invasion. (= 4 bubbles) (a strong go player told me: both are boring to play) -19x19 is critical, just in between, that's why it's fun (=3.1415 bubbles ;) 5/ Percolation: I tend to think of some dynamical systems (like spin-glasses) as naturally moving toward a static end-state where every cell is frozen (e.g up or down, black or white). (This is generally a good property for go games to have too.) But some systems just keep going. As you bring water to a boil, first you get tiny bubbles too widely spread to intact; later they start to merge; there is an edge of chaos/complexity region; and then you wind up with a chaotic boiling mess. If you removed any go rules against suicide or eye filling and made passing illegal (any empty space is legal-although it might be suicide), then a playout game would boil away forever. Just what my go engine needs ;) By tuning the playout rules, you might get different scaling effects. - Dave Hillis -Original Message- From: [EMAIL PROTECTED] To: computer-go@computer-go.org Sent: Fri, 23 Feb 2007 12:03 AM Subject: Re: [computer-go] Big board, ++physics Le jeudi 22 février 2007 01:16, David Doshay a écrit : It is pretty clear to me that, if the analogy to MC simulations in magnets is of any value, the temperature of the Go game you show is hotter than optimal. If the temperature were at the transition temperature, then each of the renormalized lattices would look just like a piece that size cut from the original. Because the details all get smaller, the original lattice is on the random, or hotter, side of the transition. Thank you very much for this work. I am mulling this over ... how to cool the Go simulation slightly from the pure MC that you did. Your analogy with physics encourage me to share other physical analogies. 1/ Cooling the simulation could be done by controlling the mixing rate and the density of stones. -Beginners'games are too cold, not enought mixed (=overconcentrated or very high viscosity, nearly solid state, not ignitable) -Professionnal games are probably near critical state (explosive conditions, gaz state) -MC-players are nearly random = too hot, too mixed, plasma state. 2/ Soap Bubbles = potential territory In addition to previous fluid state, i see hypothetical bubbles: - beginners makes some (less than 10) big bubbles, and their size and place are early known. (still too cold and too high viscosity) - professional can makes lots of bubbles (20+), but they are changing and turning very often and quickly - nearly-random makes a foam 3/ Solidification and cristal growth often comes to mind. Cristal growth need a seed to begin, generally it is a defect or some impurity. In go the defect are the corners: - they need less material to build a frontier (like soap bubbles) so corners are the beginning of the process of solidification or cristal growth. - the topology of the corner (2 libs, 3 libs and 4 libs) imposes the size and shape of a living group. - impurity is a captured stone/group 4/ shape/size resonance (un)fortunately the 19x19 size is just the critical size to have problems. -17x17 is too small, corners influence is too strong, it is quickly possible to take
Re: [computer-go] Big board, ++physics
But note that I said passing was disallowed so he would have no choice. In a percolation simulation, you can have quasi-stable regions. To follow the analogy for go, you could have situations where each color was trading small regions back and forth, like trading kos or 2 for 1 trades. Even with super-ko enforcement, you could still have some very long sequences where most of the board was stable. In principle, you could tune the playout rules to get the dynamic properties that you want. - Dave Hillis -Original Message- From: [EMAIL PROTECTED] To: computer-go@computer-go.org Sent: Fri, 23 Feb 2007 4:52 PM Subject: Re: [computer-go] Big board, ++physics This looks like the only plausible precondition: given a board of n points, n-1 are filled with the same color, and the opposing player plays the nth point, capturing the lot. Hopefully, any player of modest skill would not fill the penultimate eye of his own group. Terry McIntyre From: Chris Fant [EMAIL PROTECTED] I wonder if a large board would ever boil away to a single stone. Don't pick lemons. See all the new 2007 cars at Yahoo! Autos. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
Le jeudi 22 février 2007 01:16, David Doshay a écrit : It is pretty clear to me that, if the analogy to MC simulations in magnets is of any value, the temperature of the Go game you show is hotter than optimal. If the temperature were at the transition temperature, then each of the renormalized lattices would look just like a piece that size cut from the original. Because the details all get smaller, the original lattice is on the random, or hotter, side of the transition. Thank you very much for this work. I am mulling this over ... how to cool the Go simulation slightly from the pure MC that you did. Your analogy with physics encourage me to share other physical analogies. 1/ Cooling the simulation could be done by controlling the mixing rate and the density of stones. -Beginners'games are too cold, not enought mixed (=overconcentrated or very high viscosity, nearly solid state, not ignitable) -Professionnal games are probably near critical state (explosive conditions, gaz state) -MC-players are nearly random = too hot, too mixed, plasma state. 2/ Soap Bubbles = potential territory In addition to previous fluid state, i see hypothetical bubbles: - beginners makes some (less than 10) big bubbles, and their size and place are early known. (still too cold and too high viscosity) - professional can makes lots of bubbles (20+), but they are changing and turning very often and quickly - nearly-random makes a foam 3/ Solidification and cristal growth often comes to mind. Cristal growth need a seed to begin, generally it is a defect or some impurity. In go the defect are the corners: - they need less material to build a frontier (like soap bubbles) so corners are the beginning of the process of solidification or cristal growth. - the topology of the corner (2 libs, 3 libs and 4 libs) imposes the size and shape of a living group. - impurity is a captured stone/group 4/ shape/size resonance (un)fortunately the 19x19 size is just the critical size to have problems. -17x17 is too small, corners influence is too strong, it is quickly possible to take the border. (= 3 bubbles) -21x21 is too wide, it is not possible to quickly prevent easy invasion. (= 4 bubbles) (a strong go player told me: both are boring to play) -19x19 is critical, just in between, that's why it's fun (=3.1415 bubbles ;) I made very slow progress to formalize this, except density which is rather trivial, and a kind of temperature, but it needs a lot of go knowledge to work (something like gnugo internals), so it is not (yet) very suitable for a fast MC simulator. But the whole stuff is rather coherent in my mind. Alain ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Big board, ++physics
At 09:03 PM 2/22/2007, you wrote: 4/ shape/size resonance (un)fortunately the 19x19 size is just the critical size to have problems. -17x17 is too small, corners influence is too strong, it is quickly possible to take the border. (= 3 bubbles) -21x21 is too wide, it is not possible to quickly prevent easy invasion. (= 4 bubbles) (a strong go player told me: both are boring to play) -19x19 is critical, just in between, that's why it's fun (=3.1415 bubbles ;) don't know about bubbles, but i am under the impression that at 17x17, there corner and side territory is too large and the reverse is true at 21x21. at 19x19, there is a little less turf in in center. it's also hard to see why 21x21 would be boring (i can see 17x17 being too simple in some sense). thanks --- vice-chair http://ocjug.org/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/