[computer-go] Re: Move Prediction and Strength in Monte-Carlo Go
Hello. I updated my master thesis (http://ark.qp.land.to/main.pdf). This is a final version. I added a experiment, and I increased the number of matches with GNUGo to 600. It makes the conclusion more certain. My English does not be corrected, sorry. I will practice writing English. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Move Prediction and Strength in Monte-Carlo Go
荒木伸夫 wrote: I have considered this, and I think that this may be caused by wrong training model. In my master thesis, I mentioned that the relationship between top 1 accuracy of move prediction and the strength of Monte-Carlo is not simple (I increased the number of matches to 600, and similar tendency appeared). Therefore, it might be wrong to use only one human move (top 1 move) as a positive example (such training will highten top 1 accuracy). We may need to use another training model... Unfortunately, I don't believe a usable training model exists, besides playing plenty of games with the full MC tree search to figure out which weights produce the best playing strength. A big problem is the sample distribution. Whatever patterns we use, they are general rules with exceptions. That is to say it is always possible to make up a weird (or not so weird) position where patterns fail. And when a MC program is using patterns, it is naturally attracted towards positions that are evaluated wrongly. Rémi ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Move Prediction and Strength in Monte-Carlo Go
Rémi Coulom wrote: 荒木伸夫 wrote: I have considered this, and I think that this may be caused by wrong training model. In my master thesis, I mentioned that the relationship between top 1 accuracy of move prediction and the strength of Monte-Carlo is not simple (I increased the number of matches to 600, and similar tendency appeared). Therefore, it might be wrong to use only one human move (top 1 move) as a positive example (such training will highten top 1 accuracy). We may need to use another training model... Unfortunately, I don't believe a usable training model exists, besides playing plenty of games with the full MC tree search to figure out which weights produce the best playing strength. A big problem is the sample distribution. Whatever patterns we use, they are general rules with exceptions. That is to say it is always possible to make up a weird (or not so weird) position where patterns fail. And when a MC program is using patterns, it is naturally attracted towards positions that are evaluated wrongly. This all (combined with the results of the study) makes me think mogo and probably the other UCT programs should be searching a little wider at long time controls.At normal levels they are probably very well balanced. - Don Rémi ___ 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/
[computer-go] Re: Move Prediction and Strength in Monte-Carlo Go
Hello, Remi Coulom. Also, even worse than that, for a given set of features, the pattern urgencies computed by MM are not optimal. That is to say, it is possible to manually tweak urgencies and get a stronger program. I have considered this, and I think that this may be caused by wrong training model. In my master thesis, I mentioned that the relationship between top 1 accuracy of move prediction and the strength of Monte-Carlo is not simple (I increased the number of matches to 600, and similar tendency appeared). Therefore, it might be wrong to use only one human move (top 1 move) as a positive example (such training will highten top 1 accuracy). We may need to use another training model... ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Move Prediction and Strength in Monte-Carlo Go
Hi, I would like to confirm your experiments: I have noticed already that adding shapes of radius 4 improves prediction a lot, but does not improve playing strength (from progressive widening). Also, even worse than that, for a given set of features, the pattern urgencies computed by MM are not optimal. That is to say, it is possible to manually tweak urgencies and get a stronger program. So, as Gian-Carlo puts it, optimizing a Go program is still black magic. There is no way to avoid playing games to measure playing strength. Rémi ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Re: Move Prediction and Strength in Monte-Carlo Go
Hello. Hi, I would like to confirm your experiments: I have noticed already that adding shapes of radius 4 improves prediction a lot, but does not improve playing strength (from progressive widening). I have not yet tuned progressive widening. This information is helpful for my experiments from now. Also, even worse than that, for a given set of features, the pattern urgencies computed by MM are not optimal. That is to say, it is possible to manually tweak urgencies and get a stronger program. Oh, really!? This is great information. I'll try. So, as Gian-Carlo puts it, optimizing a Go program is still black magic. There is no way to avoid playing games to measure playing strength. O.K. I'll pay attention to this information. Thank you very much for much information. Nobuo Araki ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Re: Move Prediction and Strength in Monte-Carlo Go Program
荒木伸夫 wrote: Hello, Coulom. I'm Nobuo Araki. Thank you for reading my thesis. However, this thesis is first version, not final version. Therefore, there are too few experiments. And Mr. Hideki Kato sent me many warnings about this thesis, for example English is too bad. You may be confused while reading my English...sorry. Anyway, thanks again. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ Hi, Sorry for announcing too early. Your English is maybe a bit exotic, but not too difficult to understand. I appreciate your effort to write in English. I was very frustrated in Hakone with all those papers in Japanese that looked so interesting. Also, I believe it is not such a bad idea to release preliminary versions to the Go-programming community before producing a final version. I did it with my previous computer-go papers, and the feedback I got here helped me to improve the final version a lot. Rémi ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Re: Move Prediction and Strength in Monte-Carlo Go
Hello. Hi, Sorry for announcing too early. Your English is maybe a bit exotic, but not too difficult to understand. I appreciate your effort to write in English. In the computer science course that I belong to, we have to write master thesis in English (even graduation thesis). Also, I believe it is not such a bad idea to release preliminary versions to the Go-programming community before producing a final version. I did it with my previous computer-go papers, and the feedback I got here helped me to improve the final version a lot. Yes. I have already gotten some advice, and in the final version, I will be able to write more good paper. Thank you for commenting to my master thesis. Nobuo Araki ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Re: Move Prediction and Strength in Monte-Carlo Go Program
Hello, Coulom. I'm Nobuo Araki. Thank you for reading my thesis. However, this thesis is first version, not final version. Therefore, there are too few experiments. And Mr. Hideki Kato sent me many warnings about this thesis, for example English is too bad. You may be confused while reading my English...sorry. Anyway, thanks again. ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Re: Move Prediction and Strength in Monte-Carlo Go Program
Hi Rémi and all, It's not final version of his thesis, rather it has some (or a lot of :) errors. Please wait for the final version. -Hideki Rémi Coulom: [EMAIL PROTECTED]: Hi, I found the Master Thesis of Nobuo Araki is available online: http://ark.qp.land.to/main.pdf Abstract: Recently in the Go program, there was a breakthrough by the Monte-Carlo method using a game tree search method called UCT (UCB applied to trees, UCB stands for Upper Confidence Bounds) in combination with the reduction of search space by move prediction. By this method, Go programs easily become stronger than existing programs. However, there are hardly any studies concerning the relationship between the strength of a program, and the accuracy of move prediction, which is integrated into the Monte-Carlo method; therefore, we cannot assume the direction of future research that makes stronger programs. In this study, we developed a move prediction system based on machine learning techniques, and researched the relationship between the accuracy of move prediction, and the strength of Monte-Carlo method. Our move prediction system based on the maximum entropy method attained top level accuracies of those days. Furthermore, it became clear that even when the move prediction accuracy goes higher, the programs do not always become stronger. We investigated the reasons behind this result. Additionally, we have attempted to create a Go player by enforcing move prediction, but the result was not beyond satisfactory. We will also describe the reasons behind this result. Rémi ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ -- [EMAIL PROTECTED] (Kato) ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
