This google search provides a lot of links to sites and forums about computer shogi.
I prefer the Chess Thinkers Forum
A place for chess players to speak about what interests them:
chess or otherwise.
PS: Anything related to chess, even if it is from another Galaxy. Do you sometimes think that we were brought to planet Earth from different Planets from another Galaxy long time ago just to see how people from different planets develop as an experiment by the extraterrestrials
Last edited by pichy on Sat May 05, 2012 9:50 am, edited 5 times in total.
This google search provides a lot of links to sites and forums about computer shogi.
I prefer the Chess Thinkers Forum
A place for chess players to speak about what interests them:
chess or otherwise.
PS: Anything related to chess, even if it is from another Galaxy. Do you sometimes think that we were brought to planet Earth from different Planets from another Galaxy long time ago just to see how people from different planets develop as an experiment by the extraterrestrials
pichy wrote:
So Mr.Yonenaga Vs Bonkras is compared to David Levy vs Deep Thought 1989, in that case we are getting very close to conquer the best Shogi player.
From 1989 to 1997 is 8 years. So we are 8 years away or 2020, which was my guess when a computer will beat the world Shogi champion.
Sorry I don't agree with you it is much closer than what you think. Bonkras using parallel clusters will defeast the current champion either next year or the following like Larry Kaufman predicted.
Do I have this right? Larry Kaufman and you, Jorge Pichard, both say before 2 years or May 6, 2014 a computer will beat 1 of the top 3 professional (GM) players Watanabe, Habu or Moriuchi in a 3 hour game. I guess if any other player beats the computer then your prediction was not correct.
pichy wrote:
So Mr.Yonenaga Vs Bonkras is compared to David Levy vs Deep Thought 1989, in that case we are getting very close to conquer the best Shogi player.
From 1989 to 1997 is 8 years. So we are 8 years away or 2020, which was my guess when a computer will beat the world Shogi champion.
Sorry I don't agree with you it is much closer than what you think. Bonkras using parallel clusters will defeast the current champion either next year or the following like Larry Kaufman predicted.
Do I have this right? Larry Kaufman and you, Jorge Pichard, both say before 2 years or May 6, 2014 a computer will beat 1 of the top 3 professional (GM) players Watanabe, Habu or Moriuchi in a 3 hour game. I guess if any other player beats the computer then your prediction was not correct.
Of course we are talking about a match, not just a single game, since almost anything can happen in one game. Unfortunately the odds that such a match will actually be played are slim, given the ridiculous sums of money being demanded for a titleholder to play such a match.
pichy wrote:
So Mr.Yonenaga Vs Bonkras is compared to David Levy vs Deep Thought 1989, in that case we are getting very close to conquer the best Shogi player.
From 1989 to 1997 is 8 years. So we are 8 years away or 2020, which was my guess when a computer will beat the world Shogi champion.
Sorry I don't agree with you it is much closer than what you think. Bonkras using parallel clusters will defeast the current champion either next year or the following like Larry Kaufman predicted.
Do I have this right? Larry Kaufman and you, Jorge Pichard, both say before 2 years or May 6, 2014 a computer will beat 1 of the top 3 professional (GM) players Watanabe, Habu or Moriuchi in a 3 hour game. I guess if any other player beats the computer then your prediction was not correct.
Of course we are talking about a match, not just a single game, since almost anything can happen in one game. Unfortunately the odds that such a match will actually be played are slim, given the ridiculous sums of money being demanded for a titleholder to play such a match.
But you also predict that after May 6, 2014 a computer would always win a match with any Shogi player. Can a match be 2 or 3 games? What about the second DenOu-sen, a team match of 5 professional players versus 5 computers? The 5 strongest computers are playing in that team match. Would you agree the 5 strongest computers would always beat any 5 professional players?
pichy wrote:
So Mr.Yonenaga Vs Bonkras is compared to David Levy vs Deep Thought 1989, in that case we are getting very close to conquer the best Shogi player.
From 1989 to 1997 is 8 years. So we are 8 years away or 2020, which was my guess when a computer will beat the world Shogi champion.
Sorry I don't agree with you it is much closer than what you think. Bonkras using parallel clusters will defeast the current champion either next year or the following like Larry Kaufman predicted.
Do I have this right? Larry Kaufman and you, Jorge Pichard, both say before 2 years or May 6, 2014 a computer will beat 1 of the top 3 professional (GM) players Watanabe, Habu or Moriuchi in a 3 hour game. I guess if any other player beats the computer then your prediction was not correct.
Of course we are talking about a match, not just a single game, since almost anything can happen in one game. Unfortunately the odds that such a match will actually be played are slim, given the ridiculous sums of money being demanded for a titleholder to play such a match.
But you also predict that after May 6, 2014 a computer would always win a match with any Shogi player. Can a match be 2 or 3 games? What about the second DenOu-sen, a team match of 5 professional players versus 5 computers? The 5 strongest computers are playing in that team match. Would you agree the 5 strongest computers would always beat any 5 professional players?
Well, the "5 strongest computers" is a vague concept. If five slightly varying versions of the best program are each run on a large cluster, then they should be favored against the top five players two years from now. But if they require five unrelated programs, or if only ordinary PCs are available for most or all of them, it could be another story. I don't know if there are five super-strong unrelated programs; my impression is that most of the top ones are derivatives of Bonanza. We need details. Any info on when this match is planned to be held?
lkaufman wrote:
Well, the "5 strongest computers" is a vague concept. If five slightly varying versions of the best program are each run on a large cluster, then they should be favored against the top five players two years from now. But if they require five unrelated programs, or if only ordinary PCs are available for most or all of them, it could be another story. I don't know if there are five super-strong unrelated programs; my impression is that most of the top ones are derivatives of Bonanza. We need details. Any info on when this match is planned to be held?
The 5 strongest programs are the top 5 winners from the just completed 2012 CSA tournament. The top 5 programs in winning order are GPS, Puella, Tsutsukana, Ponanza and Shueso. Kunihito Hoki, the author of the original Bonanza entered his latest version of Bonanza. Bonanza won 9th place in the upper division contest. So Bonanza did not even play in the final competition. Bonanza ran on cluster of 6 computers with 47 processors made up of 288 cores. GPS, the winner, is open source and not derived from Bonanza. The teachers and students of the University of Tokyo, Graduate School of Arts and Sciences wrote GPS. Only 2nd place Puella has code from Bonanza. The Japan Shogi Association gave few details about the 2nd DenOu-sen competition. We know the top 5 programs from the just completed 2012 CSA tournament will play 5 pros in 2013. 1 of the 5 pros is 4 dan Funae Kouhei.
lkaufman wrote:
Well, the "5 strongest computers" is a vague concept. If five slightly varying versions of the best program are each run on a large cluster, then they should be favored against the top five players two years from now. But if they require five unrelated programs, or if only ordinary PCs are available for most or all of them, it could be another story. I don't know if there are five super-strong unrelated programs; my impression is that most of the top ones are derivatives of Bonanza. We need details. Any info on when this match is planned to be held?
The 5 strongest programs are the top 5 winners from the just completed 2012 CSA tournament. The top 5 programs in winning order are GPS, Puella, Tsutsukana, Ponanza and Shueso. Kunihito Hoki, the author of the original Bonanza entered his latest version of Bonanza. Bonanza won 9th place in the upper division contest. So Bonanza did not even play in the final competition. Bonanza ran on cluster of 6 computers with 47 processors made up of 288 cores. GPS, the winner, is open source and not derived from Bonanza. The teachers and students of the University of Tokyo, Graduate School of Arts and Sciences wrote GPS. Only 2nd place Puella has code from Bonanza. The Japan Shogi Association gave few details about the 2nd DenOu-sen competition. We know the top 5 programs from the just completed 2012 CSA tournament will play 5 pros in 2013. 1 of the 5 pros is 4 dan Funae Kouhei.
Assuming that the computers can enter their then-current version and not the one from this tournament, I would expect the computers to win the match, as it does not sound like the pros will be represented by the titleholders.
lkaufman wrote:
Assuming that the computers can enter their then-current version and not the one from this tournament, I would expect the computers to win the match, as it does not sound like the pros will be represented by the titleholders.
Did you look at the games of the winner, GPS? They are at http://homepage.mac.com/junichi_takada/wcsc22/ . I read comments critical of GPS’s games. I think the computers are not ready to beat the pros.
