Slashdot Mirror
Linux Chess Supercomputer Overpowers Grandmaster
Capt Bubudiu writes "Deep Blue vs. Kasparov is something most readers will remember but when Deep Blue was retired by IBM, a Dubai company took over with Hydra.
In a $150,000 6-game challenge in Wembley UK, the
games got off to a humiliation for mankind as Michael Adams, the
UK Grandmaster, was mauled in games one and three, drawing game two. Adams is ranked seventh
in the world and what ordinary mortals call a 'Super Grandmaster'."
The computers are not beating the humans with math. They are not relying on superior computational capability. The computers are winning with superior algorithms. Even a fairly shallow complete traversal of the search space is many orders of magnitude away from being possible, and a machine using this approach will be consistently beaten by even middling players. Computer chess has advanced primarily due to algorithmic optimizations. The evaluation functions that a modern chess engine uses are extremely well-tuned, and while a chess engine may be backed by an enormous pre-computed opening book, this too is dependent on algorithmic advances, because the book is calculated using algorithms as well, not a brute-force search. The two sets of algorithms are different, and the opening book can benefit from hugely greater computational resources, but ultimately the search algorithm is the limiting factor.
In short, the recent successes of machine chess are due to human enginuity, to the same sort of creative processes that humans themselves use to play chess. Technology, in the machine sense, is almost irrelevant (see Fritz's victories on a dinky 8P Xeon with a few gig of RAM) when compared to the advances in understanding of the game of chess.
Interestingly, even as the programmers are developing an ever-greater understanding of chess, chess players are developing an ever-greater understanding of both the game and the way in which computers play it, though people with much greater understanding of this than I tell me that the newest algorithms are playing a very human-like game, minimizing the effect of understanding 'computer chess' on the game.
High-speed Road Trip (18.000KPH)
This applies to software as it does to hardware.
No it doesn't. Do you know anything about how operating systems work? Which part do you think matters here? I/O? Just hook up a serial cable - I/O is built into the bios. Memory allocation? I seriously doubt this software is allocating memory on the fly. Process management? Why bother having more than one process? The operating system is completely meaningless. Unless you're saying Linux now has chess playing system calls built in.
Regardless of which operating system was used in this chess match, the sole determining factor is the hardware. Remember that Deep Blue defeated Kasparov with the more aesthetic MacOS, even though Kasparov is a more respected member of the chess community.
Linux zealots will cling to this "small victory", but software is only a means to an end.
I could beat the computer in a boxing match.
Anyways, everybody knows a pound of muscle weighs more than a pound of brains.
If all a computer can be is logic, I wonder if anyone has found a way to force a shutdown loop, to do something so illogical the computer can not continue.
I'm not sure what you mean by "all a computer can be is logic." If you mean that it can only follow logical arguments, I don't see that as a shortcoming. The way a computer brute-forces it's moves in a game is that it creates "game trees", where each node is a possible board state, and each branch is a possible move, either for them or thier opponent. It assigns values to each board based on things like how many moves it has available, how many pieces it has captured, which ones...
So in order to make the computer undefeatable, you need to look ahead quite a few moves to see what your opponent could possible do and counter it. This is EXACTLY what a high level chess player does, but they do it in a more intuitive way, rather than analyze every move, they analyze only probable moves, and even then, try to predict responses.
A well designed chess program will not enter a feedback loop, because if it is, say, looking 10 moves ahead, it will have its options already mapped out. The only way that a well designed chess program can get "stuck" is in the case of a draw. Now, you could force a poorly designed program to repeat its moves, but that would result in neither of you winning.
Chess can be reduced to a number of possible moves, it's just that there are so many possible moves that as of now, it cannot be completely solved.
b) our current best minmax alpha-beta game-playing algorithms are horribly suboptimal.
We know (b) is the case. Human chess players are able to "prune" much more effeciently than computer players especially in terms of eliminating bad lines. Humans are capable of much more complex "chunking" calculation than our computers (i.e. I can queen the pawn in 2 tempos). Humans are able to perform much better pattern simplification (there is no threat to the queen side).
You can watch the current game live on fics (free internet chess server). It is interesting to see how Adams has adapted his strategy thoughout this series. This game, it appears (I'm not a grand master so take this with a grain of salt) that Adams traded agressively to shorten the game. At the time of this post, Adams was down a pawn (1 rook and 3 pawns to 1 rook and 2 pawns). It also appears that Adams should be able to even the material in the next couple of moves even though Adams is currently in check. Anyway, log on to www.freechess.org and ob 37 if you want to watch.
"Those that start by burning books, will end by burning men."
Try termination checking. It's been mathematically proven to be impossible.
Justice is the sheep getting arrested while an impartial judge declares the vote void.
It should be noted that the "over-the-counter chess programs" you mention are Fritz, Shredder, Junior and Chess Tiger. These are not second-rate programs by any means; versions of Fritz and Junior have tied Kramnik and Kasparov in tournaments (and beaten them in individual games). So while impressive, it's not all that surprising that they would do well.
It's hard to be religious when certain people are never incinerated by bolts of lightning.
It's the opposite: reduced time limits helps the computer. The reason is the exponential growth of the search space. Since it grows very fast, decreasing the time affects only a bit the depth of the search.
By contrast, the human has to rely more on intuition since he/she doesn't have much time to calculate properly.
Practically speaking this is an assumption that any computing application makes about the universe it modeling.
Seastead this.