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'."

Comments

"we" won?

[moz25]

The interesting thing is that in a man vs. machine fight, the tech folks can say "we won" as they assembled the machine. Is it a humiliation or triumph for mankind that it can build a machine that can defeat itself? I think it would rather be a failure for humans if mortals can defeat highly optimized machines.

Re:"we" won?

[MindStalker]

Exactly, because while chess can be a game of strategy, it can also be a game of math via analysing moves till the end of the game. Just as there may have been people who could do math faster than 1920's computers, one the computer got the upper hand there really was no turning back, and no chance that a human could be as fast at simple calculations. Chess is the same and honestly a person may occasionally win in the next few years. Soon it will be completly solved, and no human will ever beat a computer again. (Assuming an unbeatable human goes first solution isn't discovered)

Re:"we" won?

[leonmergen]

Personally, I believe that computers, at this point, can beat mankind in anything that can be mathematically explained. Chess is an example of something that you can describe in mathematics, and thus, if you throw enough computing power at it, sure it will win. You can calculate ALL the possible moves the opponent can make to win and all the actions you can do against it at any point in the game, if only you have enough computing power.

Now, since it requires a pretty big supercomputer to win from one man, in my opinion calling it a "victory" for technology is a bit too optimistic...

Re:"we" won?

[Anonymous Coward]

Even a lever is a machine that can lift more than I can unassisted.
There is no shame in being 'defeated' by a machine.

Re:"we" won?

[WolfWithoutAClause]

More, we are human, the definitive tool user on Earth. There's no shame in that.

Re:"we" won?

[ottffssent]

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.

Re:"we" won?

[NitsujTPU]

Try theorem proving.

Computers can't beat humans at that, or even hope to do so at the moment.

Re:"we" won?

[jtdubs]

This may be a theoritical truth (not that I'm conceding that), but it is certainly not a practical truth.

Example:

The number of possibly game states in the Go is over 10^150. Many orders of mangnitude higher than the number of atoms in the universe. The best Go playing computer is ranked around 5k or so, which would make it a relatively strong amateur.

Question:

Can you name something that you believe can not be explained mathematically? Do you have evidence for this? If not, then your first sentance could be accurately paraphrased as "Personally, I believe that computers, at this point, can beat mankind in anything."

Justin Dubs

Re:"we" won?

[BewireNomali]

Exactly. There is no suspense. Machines will be better at us at these things because that is their purpose. It's what they've been designed to do.

Given enough time, machines will be better than us at EVERYTHING.

To me, it isn't amazing that machines designed to excel at chess beat the best humans. It's amazing to me that humans can still beat and/or draw games with machines designed to be brute force unbeatable.

It's like Steven Hawking beating Shaq at basketball. It's amazing, be glad that you were around to witness it. Don't expect it to happen again.

Re:"we" won?

[thue]

Personally, I believe that computers, at this point, can beat mankind in anything that can be mathematically explained.

The game Go is famous for being hard for computers to play. Wikipedia says:

Although attempts have been made to program computers to play Go, success in that area has been moderate at best - development in this area has not reached the level of Chess programs. Even the strongest programs are no better than an average club player, and would easily be beaten by a strong player even getting a nine-stone handicap. This is attributed to many qualities of the game, including the "optimising" nature of the victory condition, the virtually unlimited placement of each stone, the large board size, the nonlocal nature of the Ko rule, and the high degree of pattern recognition involved. For this reason, many in the field of artificial intelligence consider Go to be a better measure of a computer's capacity for thought than chess.


Since this problem is probably exponential in some way then it would probably currently be impossible to build a computer which could beat humans, at least until somebody invents better algoritms.

Re:"we" won?

[shobadobs]

Can theorem proving be mathematically explained?

Re:"we" won?

[NitsujTPU]

Yes, there are programs called automated theorem provers that work with, ironically enough, mathematical theorems.

It can be mathematically explained just as much as chess can be.

Re:"we" won?

[NitsujTPU]

You said exactly what I wanted to say. I was, however, afraid that my explanation would be over the heads of other readers, and I'd end up in a long thread explaining the commentary.

It would be prudent to point out an interesting strategy regarding the way humans play chess against computers. Most chess engines do do a form of search, and use techniques to optimize that search. A technique that can help a player to beat these engines is to play with a strategy that keeps the most pieces on the board, reducing the depth to which that search will reach.

Re:"we" won?

[ThreeE]

Machines will never be creative. They will always suck at art.

Re:"we" won?

[Rosco P. Coltrane]

Yes, it's classic: a student oversmarting his teacher...

I'm assuming you didn't outsmart your english teacher as a student, right?

Re:"we" won?

[MyLongNickName]

Given enough time, machines will be better than us at EVERYTHING.

Except at assigning purpose. This is one thing that cannot be expressed mathematically.

And you also do not understand chess. Chess is a drawn game by default. A "perfect" player could not beat you unless you made a mistake. There are ways to play the game that focus on minimizing risk as opposed to all out win.

Take a look at players like Petrosian (world champion in mid 60's as I recall). His style was python-like. He would see to suffocate you. Then, after tying you down, would systematically destroy you. Petrosian would be much better at playing supercomputers that Kasparov every was. His style could not be brute forced with today's supercomputers... too many plies to calculate... too many fruitless branches.

But, I do agree, in a few more decades humans will never be able to score a victory against the best computers.

But who cares? It is a linear game. I do not define my worth as a human being cased on linear criteria. Kinda gets back to the "purpose" thing.

Of course, if you are a Nihilist...

Re:"we" won?

[jbolden]

Chess computers don't look at all possibilities. They use pruning algorithms to limit options. Similarly in reversi (a game that is virtually solved via. computers) computers use a scoring system. Go is beatable, Chess is just far better known and attracted far more attention early on.

Re:"we" won?

[KDR_11k]

Who says that? Until it's mathematically proven that an artificial brain cannot be as creative as a natural one "never" is unfounded.

Re:"we" won?

[KDR_11k]

Try termination checking. It's been mathematically proven to be impossible.

Re:"we" won?

[KDR_11k]

Well, I guess another factor would be that so far there aren't any big corporations trying to make a Go AI.

Re:"we" won?

[RackinFrackin]

Chess is a drawn game by default.

This is not known to be the case. Because we do not know the optimal strategy for playing chess, we can't know the outcome of a game between two perfect players.

A "perfect" player could not beat you unless you made a mistake.

If two perfect players always draw their game, then this is true; however, if the game favors white, then there's nothing that black can do--even if he knows the optimal strategy--to win the game.

Re:"we" won?

[SamBeckett]

And you also do not understand chess. Chess is a drawn game by default. A "perfect" player could not beat you unless you made a mistake. There are ways to play the game that focus on minimizing risk as opposed to all out win.

It is you who do not understand chess!! No one knows if Chess is a drawn game "by default"--many think that might be the case, but the hugeness of the game tree prevents anyone from knowing for sure.

One could also you don't understand "risk", but that's another comment.

Re:"we" won?

[suitepotato]

Is it a humiliation or triumph for mankind that it can build a machine that can defeat itself?

I'll let you know when Skynet challenges me to a TuxRacer championship.

Re:"we" won?

[leonmergen]

The game Go is famous for being hard for computers to play.

This is because the computers aren't strong enough, compared to chess. Take a look at a chessboard - at a single moment in the game, the number of possible moves are quite limited. If you compare this to a 19x19 Go board, where you can place a stone on nearly every place on the board, the number of possible moves, if you go ahead 10 turns, are /insane/.

And in additional, Go is a game that's played by a lot of intuition - there are known groups of budhists who play the game totally on intuition, who are around 3dan.

So in the end, computers will be able to break the best go player - if only you give them enough computing power. But at that point, you can't really speak of "beating" it either, but heck, that was my original point with the current chess beating.

Re:"we" won?

[MyLongNickName]

Has it been absolutely proven? No. However, there isn't a grandmaster out there who thinks it is not a theoretical draw. If you find any expert who thinks otherwise, please post a link.

Re:"we" won?

[mikapc]

Computers will never be better as musicians as humans. They may be able to reproduce sounds extremely accurately but they obviously are lacking in everything else that makes music music.

Re:"we" won?

[leonmergen]

Can you name something that you believe can not be explained mathematically? Do you have evidence for this? If not, then your first sentance could be accurately paraphrased as "Personally, I believe that computers, at this point, can beat mankind in anything."

I think that things such as intuition, psychology and communication cannot be mathematically explained.

Re:"we" won?

[MyLongNickName]

Please tell me how I do not understand risk? Risk is uncertainty of results. When a human player who cannot see the results of his actions with 100% certainty, there is risk. When a player chooses to let his opponent have counterplay in return for opening up an attack of his own, that is risk.

Petrosian played for certainty... this means he often gave up some promising lines of play in order to reduce his opponent's chances. This is called reducing risk.

And, no, chess has not been PROVEN a draw. But please find me any expert who thinks chess is a forced win for white... If there were a forced win, several hundred years of chess (in current form) would have discovered it. Chess programs would have stumbled on it. There are opening lines in chess that have been analyze 20, 30, 40 moves deep. Yet, every analysis shows black always has some recorse, or the ability to avoid the line in the first place.

If you disagree... find me any Master level player or higher that corroberates your view.

Re:"we" won?

[bcrowell]

Check out this article on what Kasparov calls "Advanced Chess," with "a human player and a computer chess program joining forces and competing as a team against other such pairs." Nobody says bike races are pointless because people have to race with the assistance of a machine; the machine is just a tool that, like all other tools, works as an extension of the human. Nobody says Scrabble is pointless just because you have to consult a dictionary to resolve disputes; the dictionary is just tool used in playing the game, because human memory is imperfect.

And of course there are plenty of other games that computers are not yet all that good at, e.g., poker and go. Poker involves so much psychology, I don't think any machine could beat good human players unless the machine was itself sentient.

Re:"we" won?

[hunterx11]

Currently computers have analyzed chess completely where only eight pieces remain on the board. While not every move is reasonable, chess is nowhere near being completely solved. How could a comptuer even store the solution to chess, when there are more possible positions than atoms in the universe? While the time is coming soon when computers will be categorically better than humans at chess, it has not yet happened. The matches of Deep Fritz vs. Kramnik and Deep Junior vs. Kasparov both demonstrated that humans understand chess better, but are more easily fatigued and psychologically vulnerable than computers. After all, consider that a computer has to analyze tens of millions of more moves than a human to get comparable results.

Re:"we" won?

[Insipid Trunculance]

Given enough time, machines will be better than us at EVERYTHING.

How can a machine better a mother's love?

Re:"we" won?

[BewireNomali]

EXACTLY!!!! I agree with you completely. Who cares? That's the really interesting thing the AI argument investigates, this idea of human self-worth through attributes as opposed to mere existence and uniqueness. Our society and economic structure isn't geared toward a more progressive way of thinking about the individual, and about humans on a whole.

Re:"we" won?

[Anonymous Coward]

Anything called art sucks anyway, so no change there.

Re:"we" won?

[AndrewSmith1969]

Godel's Incompleteness Theorem states that in any given logical system, there are propositions which can be neither proved nor disproved. As such, while I have no idea what answer could be given to your question, I am confident that there is one. I also find it ironic that he used mathematics (set theory, IIRC) to prove that there are things that can't be proved, even though true... Andrew (insert witty .sig comment here)

Re:"we" won?

[blue trane]

Maybe you're the one who's ultimately useless to society, trying to protect what little advantage you think you have now by debunking technical progress

Re:"we" won?

[MyLongNickName]

And that is the real point. Is our "self worth" simply a figment of our imagination, religous or otherwise? Or do we really have a worth beyond our ability to produce measurable things?

And what happens when everything measurable CAN be done by a machine? Will society adapt? Will a small minority sit on a vast empire, while the majority have nothing? Or will humanity disappear entirely?

The next century should prove fascinating.

(and not only because chess will be PROVEN a theoretical draw by the end of it )

Re:"we" won?

[blue trane]

maybe all that other stuff is human-specific and arbitrary and essentially uninteresting?

Re:"we" won?

[mikapc]

yeah if you're a hardcore geek loser who is nothing more then a soulless machine.

Re:"we" won?

[WalksOnDirt]

The best results I am aware of only solve chess for up to five pieces on the board, although most six piece positions are known. These are actually for a slightly limited version of chess, as possible castling is not considered.

It may just be possible to physically store a weak solution for chess, in the far future. To prove a draw for black, only the correct responses for all possible white moves have to be stored. Similarly for the white side, of course. It's hard to make a good estimate of how much storage this would take but, at one atom per bit, perhaps a medium sized asteroid would do. Still looks too hard for anyone to actually bother doing it, even if it does become possible.

Re:"we" won?

[RackinFrackin]

I'm unfamiliar with the expert opinions, but if most or all of the grandmasters belive that the outcome is a draw then I'd believe that they are probably right. However, mathematically speaking, if it hasn't been proven then we can't claim one way or the other--we can only make a conjecture.

Re:"we" won?

[MyLongNickName]

Agreed :)

I am reminded of an International Master (I think it was Edmar Mednis) when setting up the board to the starting position: "We have now arrived at an interesting position. While it is a theoretical draw, it is very difficult to prove over the board."

This comes from my poor memory... if someone has a more accurate quote, I'd appreciate it ;)

Re:"we" won?

[mav[LAG]]

So in the end, computers will be able to break the best go player - if only you give them enough computing power.

Go is not practically solvable by throwing computing power at it, mainly because there aren't atoms enough in the observable universe to construct the computers to do the job.

Computer Go players will only get better if there's some kind of breakthrough in traditional AI or learning algorithms. Neither seems likely in the near future.

Re:"we" won?

[Paradise Pete]

Has it been absolutely proven? No.

It hasn't even been partly proven.

there isn't a grandmaster out there who thinks it is not a theoretical draw.

Not thinking it's not a draw is not nearly the same thing as thinking it's a draw. And I'll bet there are a bunch who think it's probably a draw, but are not certain, as that would be the reasonable position to hold.

And by the way (to stray from this topic to that of the match) Adams played like crap in those games. And I'll bet lots of grandmasters would agree with that.

Re:"we" won?

[BewireNomali]

I'm reminded of something from one of the Star Trek flicks, one of the ones with the borg. In it, Picard is about to be assimilated, and the Borg woman says: "I will add your distinctiveness to my own", or something like that.

It's this idea of distinctiveness as the quality with value (disturbiing in our environment of assimilationist culture) in the universe. The unique.

On the other side, there are arguments that life cannot survive its own technology. People against projects like SETI argue that it's a waste of time. They don't think any civilization has ever survived its own machines (where are the machines then?) There is no life as we know it on the other side of that event horizon, which is why we've never heard from them. They'd be telling us that our civilization, by its very inception, is doomed. Other than that, because there wouldn't be much to talk about, they've passed on making that call to talk to us.

As far as chess, I can't help but feel that white has this inherent advantage that cannot be taken from the game. It can be overcome, but to overcome something is to acknowledge an inferior position or disadvantage to begin with.

Re:"we" won?

[Monte]

When the technology becomes such that no human can beat a computer at chess (and I recon that'll be any minute now), a large number of folks will go "Oh. Well. Computers can now beat humans at chess."

And then they'll simply lose interest. I don't think there's any reasonable person that would argue that computers will never beat humans at chess - it's just a matter of "when". And when that question is answered, Ho Hum, let's go play chess with a person, because that's fun.

A computer can certainly spell better than a human. Yet that hasn't stopped spelling bees :)

And I think Reversi (Othello) has been "solved", but that doesn't mean people have stopped playing it.

And so on, and so on...

Re:"we" won?

[hunterx11]

To be fair, only a quite limited set of 7- and 8-piece endgames have complete tablebases, but I'm fairly certain that all 6-piece tablebases have been completed.

Re:"we" won?

[SamBeckett]

The game is by no means random. The game's moves cannot be modeled accurately by any "distribution", etc. (See ICC user morph2). Since the game is not random, there can be no probability assigned to it (meaningful, anyway; e.g., saying there is a 5% chance of the winning move Qxf7# is bogus). Since there is no probability, there is no risk.

And, no, chess has not been PROVEN a draw. But please find me any expert who thinks chess is a forced win for white... If there were a forced win, several hundred years of chess (in current form) would have discovered it. Chess programs would have stumbled on it. There are opening lines in chess that have been analyze 20, 30, 40 moves deep. Yet, every analysis shows black always has some recorse, or the ability to avoid the line in the first place.

If you disagree... find me any Master level player or higher that corroberates your view.

Find your own damn Master. I already said that most people think the game is a draw--why do you want me to restate it? I am simply saying that basing your logic off of "chess => draw" is premature.

Re:"we" won?

[MyLongNickName]

Non-Random does not mean "no risk". Risk is defined as the perception from the participant. If the chain of events is beyond the participants ability to calculate, then to him it is risk.

Re:"we" won?

[Ambush Commander]

Not necessarily. For example, Go is a popular game in East Asia, but success at making computers that can play go "is moderate at best".

Go is a perfect example of a game that needs intelligence, not brute force. Of course, that's not to say Chess doesn't need intelligence, but it's to a lesser extent. Now, whether or not computers will get as "smart" as us is debatable.

By the way, you should read the Wikipedia article on Go. It's really good and has been featured before.

Re:"we" won?

[NitsujTPU]

Oh, nonsense.

Here's a collection of proofs that P=NP.

http://www.win.tue.nl/~gwoegi/P-versus-NP.htm

Ooops, it seems that there are proofs that P!=NP in that collection as well. Maybe you're on to something.

Re:"we" won?

[Fortress]

Except at assigning purpose. This is one thing that cannot be expressed mathematically.

Maybe there are new mathematics that we haven't discovered yet.

Re:"we" won?

[Deliveranc3]

Really, since every grandmaster has to study what came before them we build the grandmasters too.

To be scared of your student such that you would weaken them is absurd and vaguely perverted.

We knew they would be better, we knew cars would be faster. We'd better just suck it up and try and make them even better, pushing what we can imagine them doing, beating a human is great but computers have yet to beat each other to the point where the games are trivial though it shouldn't be long.

Re:"we" won?

[Mornelithe]

And I think Reversi (Othello) has been "solved", but that doesn't mean people have stopped playing it.

Hell, some people still play tic-tac-toe.

Re:"we" won?

[Fortress]

Can you name something that you believe can not be explained mathematically? Do you have evidence for this?

While I may not be able to name something, the evidence suggests that there are some true things that cannot be proven mathematically. Someone even proved this mathematically!

Re:"we" won?

[MarkRebuck]

>And you also do not understand chess. >Chess is a drawn game by default. A >"perfect" player could not beat you >unless you made a mistake. No. This is an entirely open question. My guess is that a "perfect" game of chess is a draw with KRP vs. KR (ie White has a material advantage but can't force a win). But that is only a guess. The simple fact is that we don't know, and never will know (the problem size is way, way, way, way, way beyond knowing) the value of a "perfect" game of chess. >But who cares? It is a linear game. No. It's exponential. Specifically, exponential with a branching factor of something averaging around 20 per ply. Pruning algorithms reduce the branching factor of a search to something between 3 and 4, but it is still exponential. And, for what it is worth... Petrosian was a hack. Kasparov is the best chess player (human vs. human or human vs. computer) the world have ever seen. Tal, or perhaps Fischer, had the most style, but Kasparov would beat either over the board.

Re:"we" won?

[gordo3000]

GO is not more complex than chess by any means. It is the simplicity of the game that gets computers. The problems is the sheer size of the board. Unlike chess, there are 225 opening moves(15x15 board) in Go. The patterns are actually quite simple in the game, except they can be entered into in so many different ways.

The funny thing is its hard to program computers to see patterns. Early computers brute forced chess, and they were terrible. Most human players well below the master level would be even opponents. The best programs today do a mixture of brute force and pattern recognition.

Programming Go is just not the holy grail that programming chess is right now. Give it time, and maybe IBM, and you will probably see huge leaps and bounds in the skill of computer controlled players.

Re:"we" won?

[poopdeville]

Eh, I don't think we can say who the best player the world has ever seen is. Capablanca was the best player in his time, and Kasparov might be the best now. But without at least a nice one-on-one seven round match, there's no point in comparing them.

Re:"we" won?

[MyLongNickName]

I knew you were a troll when you said "Petrosian was a hack". Even Kasparov praised him...

Calling a world champion a hack... what is your ELO?

Your logic is only surpassed by your text editing abilities...

Re:"we" won?

[WalksOnDirt]

I knew the six piece set was mostly done. Before posting I checked around for anything saying the six piece tablebases were finished, but I couldn't find anything. I check Hyatt's site and some are obviously missing, but that doesn't prove anything.

If you have link for the completion of them, I would be interested in seeing it.

Re:"we" won?

[poopdeville]

Unfortunately, a mathematical "explanation" (what an ugly term) is not sufficient for machine computation. Witness the use of the Second Order Axioms of Peano Arithmetic to prove that there is a unique model of the natural numbers. This would appear to contradict Godel's Incompleteness theorem, since if there is a unique model of the natural numbers, every sentence true relative to that model would have a proof, by the completeness theorem.

So what is the difference between the theories of the first and second order Peano Axioms for arithmetic? Well, in the first order theory, one can only quantify over so-called definable sets. These are exactly the decidable sets from recursion theory. In a second-order theory, one quantifies over all sets.

Since an automated theorem prover must be given a recursive set of axioms from which to build a theory, in any domain where the family of definable sets is a strict subset of the power set of the domain, there will be theorems that can only be proved with help from second-order logic.

Re:"we" won?

[poopdeville]

Godel's Incompleteness Theorem states that in any given logical system, there are propositions which can be neither proved nor disproved.

False. Pressburger arithmetic is complete -- which is to say that every proposition can be proven true or false. Hilbert's formulation of Euclidean geometry is also complete.

Re:"we" won?

[theapodan]

Parent says about chess "soon it will be completely solved."

Is chess a game that lends itself to be solved mathematically? I have little doubt that computers will be able to compute to such a depth as to never lose a game, but I don't know if chess is solvable, like Connect 4. http://www.connectfour.net/Files/connect4.pdf(Warn ing: It's a PDF) Or by being solved, did you simply mean that you agree that chess can be played to such a depth as to make it impossible for a human to win? That's not really "solving" a game.

Re:"we" won?

[kmhebert]

you don't understand "risk", but that's another comment.

Isn't that the game with the cannons?

Re:"we" won?

[kmhebert]

Hell, some people still play tic-tac-toe.

| |
---
|X|
---
| |


Your move!

Re:"we" won?

[TaoJones]

Go is beatable

After 32 moves or so the number of possible states exceeds the number of particles in the universe. Compute that :)

Re:"we" won?

[ahdeoz]

Also, the simple fact is that chess players (and programmers) understand the game far better than Go players do. Go *is* an amateur game and even the best players don't really understand the algorithms that govern the game. A good Go player beats a weaker player because he is more familiar with the patterns that work and is good at identifying captures and avoiding traps. There isn't really a strategy, in the sense. It's more about feints and cautious defense.

Re:"we" won?

[ahdeoz]

There are more possible moves in Go than chess, so it is a better measure of a computer's capacity for thought? If the emphasis is on "capacity", okay, but what exactly do you mean by "thought"? Most people don't consider thought to be defined by quantity of decisions.

Re:"we" won?

[amehra]

I believe that computers (machine?) would never come close to humans in things that call for creativity: Composing music is an excellent example. Also poems, drawing paintings and such. I don't think computers could ever come close to matching our best composers, poets and painters.

Re:"we" won?

[Jedi Alec]

Use quantum physics to drag in a couple of parallel universes to help out with the calculating :P Then when we have finally created the ultimate quantum-chess computer to "solve" chess all of creation folds in on itself while softly in the background we hear *insert deity of choice* saying "checkmate".

Re:"we" won?

[isomeme]

I believe it was Marvin Minsky who, when asked whether computers would ever be as smart as humans, replied:

"Yes. Briefly."

That one always gives me the shivers.

Re:"we" won?

[julesh]

if most or all of the grandmasters belive that the outcome is a draw then I'd believe that they are probably right

The community is split. The fashion of the mid-late 20th century was to think that it was a draw. Opinion is now swinging towards the "perfect game" being a win for white.

Unfortunately, it's likely that no definitive proof will ever be produced, as the problem space is simply too large to be calculated effectively, and no provably correct shortcut has yet been found (despite much effort).

Re:"we" won?

[julesh]

A technique that can help a player to beat these engines is to play with a strategy that keeps the most pieces on the board, reducing the depth to which that search will reach.

I also believe that playing with reduced time limits also favours human players, as it increases the amount of intuition involved in assessing situations.

Re:"we" won?

[julesh]

Go is not practically solvable by throwing computing power at it, mainly because there aren't atoms enough in the observable universe to construct the computers to do the job.

The same is true for chess, which has about 10^40 possible positions, which is I believe beyond the theoretical limit of calculability which has been posted on slashdot before.

Re:"we" won?

[mav[LAG]]

You're right. I meant to refute the GP's assertion that computers will one day beat the best Go players by brute force, not drag in a discussion about solvability :)

Re:"we" won?

[jbolden]

And if the pruning algorithms were effecient enough it wouldn't matter. For example assume that the computer were able to eliminate 99.9% of all moves. Then it would only have to examine a very small number of possible states to go all the way to end of game. Go is just an exponential function, pruning is an exponential function. Its just a question of making the denominator grow rapidly enough. Further if positional analysis becomes more possible than the the computer doesn't need to solve to the end.

Re:"we" won?

[Anonymous Coward]

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.

Re:"we" won?

[rbarreira]

| |X
-----
|X|
----
| |

Re:"we" won?

[kmhebert]

| |x
-----
|x|
----
x| |

I win!

Re:"we" won?

[NitsujTPU]

I'm sorry, you didn't read what I said correctly.

Same amount of time, keep the same number of pieces on the board.

A search algorithm will make less progress in the same space.

I'm correct on this one.

Re:"we" won?

[NitsujTPU]

My apology, I was browsing with my threshold too high and missed the parent poster.

Re:"we" won?

[rbarreira]

AHAHA best tic tac toe game ever... I'm so stupid ;)

Re:"we" won?

[coopex]

Could it be impossible for humans (ie mathematically, or beyond the capacity of the brain) as well?

Re:"we" won?

[MindStalker]

Unfortunately, it's likely that no definitive proof will ever be produced, as the problem space is simply too large to be calculated effectively, and no provably correct shortcut has yet been found (despite much effort).

Watch that word "ever". Computers are becomming faster all the time, certanly SOMEDAY chess will be solved. (Unless of course civilization is destroyed first)

Re:"we" won?

[MyLongNickName]

Sounds like a "team building event" at management training school: Cooperative tic-tac toe. You and your teammate have to get three in a row. Alternating, each of you puts an X on a three by three grid. Can you finish in six moves or less?

Heck, I'd bet three out of four of them could do it!
.
.
.
.
.
.
.

(and yes, I do know it is impossible NOT to do it in six moves)..

Re:"we" won?

[julesh]

Computers are becomming faster all the time, certanly SOMEDAY chess will be solved. (Unless of course civilization is destroyed first)

I don't believe so. Unless some way of circumventing the laws of thermodynamics is found, there is a theoretical limit on the maximum amount of information that can be processed within our universe, which has been calculated to a figure not entirely dissimilar to the order of the tree of all possible chess games.

This calculation makes a few assumptions that seem likely to hold out. The only way around it would be a quantum solution, and chess doesn't seem like the kind of problem that quantum computers can solve.

1. e4

[Anonymous Coward]

First move!

Re:1. e4

[shobadobs]

You moron. You can't make white's second move until black has made its first move! And if you look at the other reply ("1. - e5"), it looks like black already occupies that square.

I dont get it...

[__aaxwdb6741]

I dont get it. Why is it so amazing that computers beat human beings in chess? Isnt chess all about logic and calculation? Arent computers all about logic and calculation?
If both are true, then how come it is so amazing that a computer beat a human being in chess?

Wouldnt it be more amazing if a human being beat a chess computer?

Re:I dont get it...

[Hognoxious]

We could try pulling the plug ... no wait, that might provoke it into starting a nuclear war.

In which case I for one welcome our new en-passanting overlords.

Re:I dont get it...

[vortex2.71]

Because computers can only perform arithmatical operations and "playing" chess is enormously more abstract than most other current computer functions. Second, its interesting because the Kasperov vs. Deep Blue match was vehemently disputed and IBM then conspicuously took the computer offline and refused a final rematch. This ended the possibility of a best 2 out of 3 set of matches. Finally, its really damn interesting that brute force computers that can do some really amazing calculations can't reliably beat super grand masters yet... It gives us insight into the ultimate potential of computers that not only calculate, but calculate in an intelligent manner. Think about a machine that could combine Kasperov's intelligence and deductive ability with Deep Blue's raw computing power. It would be fricking brilliant!

Re:I dont get it...

[j1m 5n0w]

Why is it so amazing that computers beat human beings in chess?

...because no computer has been able to consistently beat the top chess players. It is amazing that computers that can evaluate millions of board positions per second can still lose to a human who can evaluate maybe tens of board positions per second. Go is even worse; the best go programs in the world are routinely defeated by mediocre go players. This means that either a) the human brain can do a lot more computation than we give it credit for or b) our current best minmax alpha-beta game-playing algorithms are horribly suboptimal.

Re:I dont get it...

[barcodez]

Damn, and I'm routinely beaten by mediocre go programs :(

Re:I dont get it...

[Inspector Lopez]

In the case of Go, computers perform abysmally compared to humans.

While very high-end computer chess machines now play strong grandmaster chess, it takes relatively little practice to beat the best Go-playing computer.

In chess, "search" is part of the computer algorithm, and it is hard in chess because the tree of possibilities gets big in a hurry.

But in Go it is far worse.

In chess there are (I think) 16 first pawn moves + 4 first knight moves, and the same holds for black --- so that there are 400 possible positions after white and black have moved.

In Go, however, there are 361 possible first moves, and 360 possible second moves; divide by 8 if you wish for rotational symmetries, resulting in some 16,000 different possible positions after the first round of moves.

Humans have a spectacular ability to detect patterns which has yet to be duplicated in machines --- as anyone practicing speech recognition or image understanding can tell you. It is in fact quite remarkable how well the chess machines are doing.

Re:I dont get it...

[jbolden]

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).

Re:I dont get it...

[gmarceau]

Because when people started researching artificial intelligence, chess playing was indentified as a problem requiring human-style hinsight.

So, now that compuiter can beat grand chessmasters, it forces people to reconsider what is intelligence, or, alternatively, to admit that the computer simulates human insight in some form.

Re:I dont get it...

[public transport]

Computers are in fact not that good with logic. We easily formulate the rules of chess in a state-of-the-art solvers for first order logic, and it would make a horrible chess player. This is because first (and higher) order logic is NP-complete, which means it is quite easy to come up with a problem that it can not solve within reasonable time on today's or tomorrows computers (until we have usable quantum computers).

Humans can solve logical puzzles which the computer can't, because we use intuition and expert knowledge to prune what possibilities to investigate. I know there is at least one logic system, Coq, which is designed to let a human guide the proof search, combining the precision and brute force of computers with human intuition.

Re:I dont get it...

[KDR_11k]

As long as the human you throw in there is some pasty geek who never lifted anything heavier than a can of coke...

Re:I dont get it...

[HermanAB]

Chess is a very 'large' problem. You cannot simply proram all possible outcomes and then do a brute force pattern match. For example chess programs - including Deep Blue - use a method of play based on the `state-space search' (SSS) model. How programs like this work can be illustrated using a simpler game - Noughts and crosses. Size of search space in Noughts and Crosses: # First move: 9 possibilities # Second move: 8 possibilities # Third move: 7 possibilities Total number of move sequences: 362880 (= factorial 9) Searching each of those sequences, at the rate of one per second, would take about 100 hours. This is an illustration of what is called the `combinatorial explosion'. Search spaces can become huge, even when they involve a relatively small number of alternatives at any single node in the search tree. With chess, the problem space is much, much larger - factorial 64.

Re:I dont get it...

[Frodo2002]

"Isnt chess all about logic and calculation?"

I wish to repond to this. Chess (as played by humans) is definitely not ALL about logic and calculation. It is ALSO about creativity, ingenuity and occasionally heroism. That is the beauty of the game. To be able to study a game between two GMs and be able to see and appreciate those human qualities - that is what makes it special. I don't care if computers can finally calculate fast enough to beat the best human players. Chess is a lot more than that, otherwise humans would have given up playing it long ago.

Re:I dont get it...

[KDR_11k]

Dunno, what is "mountain dew"?

I wonder...

[the linux geek]

Is it possible that a computer could compute every possible move, make a database of it, and win automatically every time?

Re:I wonder...

[Hawkxor]

In fact, i've heard that it's probably impossible for a perfect strategy in that sense to exist, as there aren't enough atoms in the universe to store the amount of data which would the computer would be required to hold.

Re:I wonder...

[Cyberax]

It's been proven that there is a perfect strategy for chess (see http://en.wikipedia.org/wiki/Game_theory), but the number of variants is astronomically large (more than a number of atoms in the whole known Universe).

However, it might be possible to create a quantum computer which computes the best strategy for 1000 turns (virtually an 'ideal' strategy). AFAIK, there are some researches on this topic.

Re:I wonder...

[tmasssey]

This is exactly the idea of "solving" a game. By analysis, you can determine how exactly a perfectly-played game will end. For example, you have most likely solved tic-tac-toe: you know that the person who goes first always wins, when they play properly, no matter *what* the other person does.

Of course, TTT is slightly simpler than chess... That's why it's not yet been solved. Given the number of legal positions on the board, it is unlikely that a game like chess will ever be solved (by some estimates there are more entries in the game tree than atoms in the universe). So possible in the theoretical and possible in the practical are two different things! :)

Re:I wonder...

[MerlynEmrys67]

Your Move - show me your guaranteed winning strategy.

Start your move in a reply

Ok - first off, you haven't even mastered Tic Tac Toe... I'll let you have the first move, and guarantee at best you will get a draw, of course if you make a mistake - I can win with the second.

Re:I wonder...

[Anonymous Coward]

Err... Not to be picky, but what you mean is atoms in the *visible* universe. Actually we don't know what is outside of the visible part of the universe and probably never will, since the expansion is accelerating and the visible part is decreasing. It's very possible that there's an infinity of atoms - enough to hold all games of chess.

Re:I wonder...

[Ki Master George]

It is possible, I suppose, but there are just too many moves and too many possibilities to do that. But somewhere out there, there is a strategy in chess that cannot be beaten, although we will probably never reach it.

Re:I wonder...

[Hussman32]

There are hash tables of the various endgame positions, and those are solved. The computer also usually has a complete list of chess openings, so they are guaranteed to open with equality or strength for the first three or four moves. Where computers can lose is the middlegame.

Re:I wonder...

[Fermatprime]

While it is conceivable, the sheer number of moves that the computer would have to calculate would put a method like this beyond all practical application. And even then the computer wouldn't ALWAYS win - what if it played itself?

Re:I wonder...

[MerlynEmrys67]

Lets say that both players don't make a mistake - my 6 year old won't loose a game of TTT - and I would hardly call her a skilled player

Re:I wonder...

[ToasterofDOOM]

No, I think there isnt a "perfect" chess strategy ... there is always a way to beat it. In game theory, optimum choice is the best possible thing you can do. tic-tac-toe suck because if you always make the optimum choice, you rule out losing. chess is far too complex to ever perfect, not just because of computing power but because chess perfection is nonexisent

Re:I wonder...

[swilver]

If Moore's law manages to hold another 100 years, computers will be fast enough to calculate all moves in a chess game in about an hour orso.

However, that wouldn't mean they would automatically win. It's more likely that they could always force a draw, and only win if the opponent makes a mistake.

Of course, Moore's law is highly unlikely to last for another century, as it is already showing signs of breaking down.

Re:I wonder...

[e2ka]

It could be like your peer poster was saying about tic tac toe as a solved game. Perhaps such a game would be reduced to whoever goes first will win.

Re:I wonder...

[MrByte420]

For example, you have most likely solved tic-tac-toe: you know that the person who goes first always wins, when they play properly, no matter *what* the other person does.

Ummm....Tic Tac Toe is always a tie under optimal play.

Re:I wonder...

[capologist]

Nobody knows how much data the computer would be required to hold. The "analysis" that "proves" it impossible proceeds from the assumption that the perfect computer would have to store a representation of every possible position.

AFAIK, it has not been disproven that there is a yet-to-be-discovered invariant, computable by real-world computational power, that determines whether any given position is a win for white, win for black, or draw.

Re:I wonder...

[Krimszon]

But if you zip it...

Re:I wonder...

[VacaBoi]

No. Chess is a problem that has challenged computer scientists since the dawn of computer science precisely because no conceivable computer could ever analyze all possible moves. Back in the 1940s, Claude Shannon at Bell Labs estimated that the total number of moves was approximately 10^120. So, computers essentially have to play chess the same way as humans. A model of the opponent is created, and the computer crunches through as many scenarios as possible in a reasonable amount of time. Ultimately, they have to anticipate the moves of their opponents based on experience and reason. This is what makes CS Chess so interesting. Chess is not, in practice, the linear problem it theoretically should be.

Re:I wonder...

[tmasssey]

You are correct. Change "win" to "tie"...

Been playing with my daughter for too long. I haven't lost a game of Tic Tac Toe in a while... I forgot that it's supposed to end in a tie! :)

I Have To Say It...

[Doc Squidly]

I, for one, welcome our new Chess Playing Computer Overloards.

Re:I Have To Say It...

[Quarters]

I really wish Slashcode would include moderator ratings for "Trite", "Cliche", "Overused", "Never was funny, and "No, really, it WAS NEVER FUNNY!"

Re:I Have To Say It...

[blue trane]

It was funny in Soviet Russia.

Re:I Have To Say It...

[bladx]

There have got to be better jokes on Slashdot than these :/

Re:I Have To Say It...

[antifoidulus]

eh, it was only funny to the old Korean Soviets

Re:I Have To Say It...

[KDR_11k]

The term you're searching for is "Redundant". If everything else fails, use Overrated.

I think...

[mogalpha]

No one cares and or has any mod points today :)

Re:I think...

[canuck57]

No one cares and or has any mod points today :)

The computer won them all.

Re:I think...

[mogalpha]

Oops, I guess I'm wrong. People do care and or have mod points today! But +2 funny? Man that's overrated. Luckily no one has the mod points to mod me down.... oh crap.

It is inevitable...

[Skiron]

... that computers will beat a man at chess all the time they are allowed to use a database on positions.

The time to get scared is when a 'thinking' computer chess program does it all for scratch from the first move.

Having said that, GNUChess 0wn35 me bigtime, the bugger.

Re:It is inevitable...

[cryptoz]

Eh? I didn't think they used databases at all, I thought they were actually doing the calculations all from scratch? Right? Did you RTFA?

Re:It is inevitable...

[Anonymous Conrad]

The time to get scared is when a 'thinking' computer chess program does it all for scratch from the first move.

From the FAQ here,

The Hydra opening book is very short. Typically 10 moves. After 10 moves we let the monster from the leash and rely on the playing strength of the program.

which is probably less than a serious chess player will have, albeit broader.

Re:It is inevitable...

[MerlynEmrys67]

Not even humans play without a database.

If you look at high end chess - you play so many moves in a certain time (20 moves in 4 hours maybe) - so to make sure you have enough time when it is interesting, you start standard moves, and until you or your opponent go outside of an silently agreed on game, the moves are fast and furious (watch the first 10 moves when 2 grand masters play) - then they slow down as the players try to figure out when to deviate from the script, then about 12-13 moves in (in some cases) the plays start taking about 20+ minutes a turn.

So yes - openning databases are known quite deeply by the best players - a computer using a database is only fair.

Re:It is inevitable...

[dvdeug]

... that computers will beat a man at chess all the time they are allowed to use a database on positions.

The time to get scared is when a 'thinking' computer chess program does it all for scratch from the first move.

If we're playing a fair game, and the human doesn't get to use a database of positions--has never played a game of chess or read about chess position--then I suspect the computer will win. I fail to see why it's unfair for a computer to have databases against a human who has studied opening books and endgames for years.

In 50 years..

[btgreat]

In 50 years will chess club be dominated by nerds who know how to build computers and write software or by the humans who take the time to learn the game? Society is becoming more and more oriented towards computers and I wouldnt be surprised if in the future people judge their skill based on who can write a better program for their computer, rather than knowing how to play the game itself. It's just too bad these computers don't give lessons.

Re:In 50 years..

[Anonymous Conrad]

In 50 years will chess club be dominated by nerds who know how to build computers and write software or by the humans who take the time to learn the game?

But the exercise here is the evolution of chess algorithms, and to produce a more natural opponent. You could draw parallels with the advancement of sport science to improve athletic performance over time - we're building better training aids.

(But let's not get carried away with the analogy: anyone who says chess is a sport is an ass.)

Re:In 50 years..

[m50d]

People will know it, like today people do with draughts or even noughts and crosses (tick-tack-toe to americans). But I don't think there'll be serious players like there are today. I've been the county champion in my age group for the past three years. The last one was for lack of competition. In the time I've been playing competitively I've seen a decrease by at least 75% in the number of people competing in junior tournaments. Make no mistake, competitive chess among humans is dying.

Re:In 50 years..

[Metasquares]

The chess club is already dominated by nerds who know how to build computers and write software :)

Re:In 50 years..

[malraid]

Let me guess, you are a "nerd" and want recognition because of it? Well, tough luck! Society is based on people, and peoples' skill are judged based on how well you can comunicate those skill to other people. Most "nerds" are very intelegent, but the lack a key ability, charisma. Work on your charisma, everything will follow. Life is like AD&D, but on the game, as on real life, most "nerds" give little importance to that skill. I know I did for a very long time. Trying to improve my charisma in real life has been one of the most gratifing things I've done.

Re:In 50 years..

[John Seminal]

In 50 years will chess club be dominated by nerds who know how to build computers and write software or by the humans who take the time to learn the game? Society is becoming more and more oriented towards computers and I wouldnt be surprised if in the future people judge their skill based on who can write a better program for their computer, rather than knowing how to play the game itself. It's just too bad these computers don't give lessons.

It is one of the reasons I hate playing chess online. There are not many, but every now and then I run into someone who is running chessmaster on their computer. They just want to win, they don't want to play the game.

Chess is a relaxing game. It is supposed to be intellectually pure and honest.

It is too bad that Vegas never started with chess tables. That would be something to do, better than blowing money on the slots. There would be no cheating. Maybe in order to play you would have to register with the casinos, and they would keep track of your wins and losses and give you a score so people would know your level.

Re:In 50 years..

[blue trane]

Just think, all that wasted energy spent getting really good at a solved game, maybe now they can apply their brain power to an as yet unsolved problem.

Re:In 50 years..

[GoldenBear]

As a nerd who write chess software for computers, I find this comment hilarious.

Re:In 50 years..

[nkh]

Work on your charisma, everything will follow.

Is there really such a thing as "improving your charisma"? From my (rather short) experience in life, I discovered that once you've mastered a subject, your work of communicating this knowledge is almost done: just speak about what you know but do it clearly (in good english or whatever you use instead) To me "charisma" is more like "technical skills + vocabulary" and I'm neither Steve Jobs nor Brad Pitt but I can talk to other people about what I do once I use carefully chosen words.

Re:In 50 years..

[m50d]

Process is as important as the solutions you finally reach. Programmers don't do hello world, towers of hanoi etc. because they're problems waiting to be solved, but to get good at programming. I think chess helps people be more analytical, and is worth learning for the mental effects on yourself (as are plenty of other similar games, like draughts or go).

Re:In 50 years..

[jbolden]

Since when is chess a solved game?

Re:In 50 years..

[jbolden]

Lots of parks in the east coast have speed chess gambling. Washington Square Park in NY City is the most famous location.

Re:In 50 years..

[blue trane]

Sure, but you don't need to become a grandmaster (i.e. spend a lot of time) on chess to benefit from it. Like those "hello world" programs, you write them once and move on. Just saying that if Kasparov or whoever knew they couldn't defeat a computer in chess, they might have dedicated their brain power to some unsolved problem, or gone into programming or something.

Re:In 50 years..

[malraid]

Let me put it this way, imagine that two people go to an interview. One is shaking from the nerves. The other one is confident, is able to engage the interviewer in some casual talk about the company, thus showing interest and knowledge in the company, and posibble the industry. Other guy just sits there avoiding eye contact, anserwing "ehhh....ahhhh...hummm....yes". Who do you think will get the job? And I'm not even talking about technical skill here. I've noticed that "about what you know but do it clearly" is a skill that a lot of people don't have, and it's a skill that they should work on improving.

Re:In 50 years..

[slashdot_commentator]

Except its not gambling in NYC. The law here defines gambling as a game of chance. There is no chance in chess, and there is no law here against betting on games of skill.

Bottom line, head out to Washington Square Park, and drop your money down again a willing opponent. No cop will arrest you for playing chess with a wager on the winner.

Jessie Owens Outpaced by Motorbike

[nagora]

Big deal.

TWW

Re:Jessie Owens Outpaced by Motorbike

[m50d]

There were always animals who could outpace us running. We never met anything else that could beat us at chess.

Re:Jessie Owens Outpaced by Motorbike

[alexo]

> We never met anything else that could beat us at chess.

So we built one.

Re:Jessie Owens Outpaced by Motorbike

[nagora]

There were always animals who could outpace us running. We never met anything else that could beat us at chess.

Firstly, there are in fact no animals that can outrun the best humans over long distances. Secondly, we've had calculators that can work out logs faster than we can for a long time; chess is no different.

TWW

Re:Jessie Owens Outpaced by Motorbike

[Dirtside]

Yeah, but we invented chess. Running was around for hundreds of millions of years before humans showed up. Is it any surprise that we're (so far) the best species at a game we invented? :)

Re:Jessie Owens Outpaced by Motorbike

[m50d]

No. It's surprising that we got beaten at it.

Re:Jessie Owens Outpaced by Motorbike

[Dirtside]

But we haven't been beaten at it. We built a tool to play it better -- it's not fundamentally different from inventing baseball, and then inventing a machine that can hurl baseballs at 500 miles per hour.

If we came across a species of dolphin that could beat us at chess, then yes, I'd say surprise was a valid reaction. But we built a machine for a specific task, and you're surprised that it's good at that task?

Re:Jessie Owens Outpaced by Motorbike

[nagora]

There are african tribes that hunt by running animals to death. I believe the animals in this case are motivated enough. :)

Correct.

Re:Jessie Owens Outpaced by Motorbike

[m50d]

It was once thought to be uniquely human - when AI research was getting started, a machine that could play chess as well as a human was thought to be a major milestone, similar to the Turing test. A computer better at playing chess than us is a step towards a computer better at being intelligent than us, and then a computer better at being human than us.

Re:Jessie Owens Outpaced by Motorbike

[nagora]

A computer better at playing chess than us is a step towards a computer better at being intelligent than us

Not at all. Chess hardware and software is now so specialised that it tells us absolutely nothing about intelligence.

TWW

Hooray?

[jackcarter]

If a computer could do it 8 years ago, then with Moore's law, this is 1/(2^5) as interesting as it was then. Did it quickly by hand.

Re:Hooray?

[WolfWithoutAClause]

Deep Blue ran on purpose designed ASICs whereas Hydra runs on FPGAs which are slower, but more flexible. Turns out that the inherent slower speed of FPGAs and Moore's law roughly cancels.

However, the chess technology in Hydra is 8 years newer in other respects, and so Hydra is able to look about 8 moves further ahead (albeit with slightly less accuracy, but it turns out it's a pretty big win anyway). So Hydra would be expected to comfortably beat Deep Blue, should they ever meet, which is unlikely in fact (although Deep Blue still exists, IBM are hardly likely to boot it up just to lose).

Face it...

[TheStick]

From now on, we are sure Humans are dumb.

Thank goodness I'm a Vulcan!

Re:Face it...

[not-enough-info]

I am a Jelly Donut!

werke

[mnemonic_]

Read about how chess computers work. There are 10^120 possible moves for a certain "tree" sequence of moves. Today's chess computers evaluate millions of moves per second, far short of all possible moves, due to computing limitations.

It's interesting to note that both grandmasters and amateurs have been shown to think only 3-5 moves in the future, while computers calculate for 10-20. Despite that, humans are still competitive with computers in chess (losing some games, winning others), showing there's more to the game than how far one can predict. Those 3-5 predictions of a grandmaster will differ from those of the amateur, and those 10-20 of the computer.

Other uninteresting things computers can do

[shobadobs]

Computers can also multiply hundred-digit integers faster than humans.

I'd like to see a computer beat the best Go players. Or how about a computer that can beat the best human chess players at Fischerandom chess

Re:Other uninteresting things computers can do

[m50d]

When computers beat top draughts players people said exactly that about chess. Go is going to fall, it's just a matter of time.

Re:Other uninteresting things computers can do

[shobadobs]

Yeah, it will. But will it happen because Go playing programs become smarter, or will it happen because computers become faster?

Re:Other uninteresting things computers can do

[m50d]

The programs will have to become smarter. Brute forcing does not work with go, you have on the order of 381 possibilities each move. 381^x gets very big very quickly. But I don't doubt the programs can get smart enough.

Re:Other uninteresting things computers can do

[starbird]

Both

Re:Other uninteresting things computers can do

[dyefade]

Surely brute force, will work, but only if computers become (unimaginably) more powerful. Wasn't that the point of the post?

Re:Other uninteresting things computers can do

[m50d]

Technically so with go as it's currently played (though I don't think we can rely on Moore's law to hold out for anything like long enough, so that doesn't make computers winning at go inevitable on its own), but it wouldn't really be a proper solution. We could make it 21x21 without altering the dynamic so much for human players, wheras that would set a brute force approach back immensely.

Re:Other uninteresting things computers can do

[dyefade]

Excellent point. Yeah, I'd actually read this whole discussion before, not sure why I didn't remember that. Cheers!

Re:Other uninteresting things computers can do

[julesh]

Brute forcing does not work with go, you have on the order of 381 possibilities each move. 381^x gets very big very quickly.

That was the same argument used with chess. There are on average about 30 possible moves in a chess position. 30^x gets big quickly enough that you can't just build the tree and hope. You have to build it selectively and use heuristics to decide what's worth investigating and what isn't.

If go is to fall, similar (but more effective) heuristics will need to be found for it. I'm not familiar enough with the game to know how easy this will be.

Re:Other uninteresting things computers can do

[m50d]

Yes, and that's why chess didn't fall to pure processing power. If you took a chess program from 20 years ago and threw (modern) processors at it, you still wouldn't get near beating grandmasters.

I still get beaten by GNU go, so I'm not sure I'm good enough to judge. But just looking at the numbers, you'll basically need to be able to eliminate 10x as many alternatives to be able to work forward as many go moves as you can with chess.

Linux?

[anthony_dipierro]

Why even mention the Operating System in something like this? It's pretty much irrelevant what operating system you're using, in fact you could probably spend two days or so converting the program to run without any operating system at all.

Re:Linux?

[daniil]

The poster is probably a Linux fanbox and assumes that, had the computer been running any other operating sytem, it couldn't have done it.

Re:Linux?

[kaleco]

Because the constitution of the machine that can achieve such a coveted triumph is very important. This is why IBM staged the original Deep Blue event against Kasparov - to raise publicity for their equipment.

This applies to software as it does to hardware.

Re:Linux?

[anthony_dipierro]

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.

Re:Linux?

[samoa]

well, i don't know anything about how operating systems work, but i'll have a crack :)

this 'computer' is a clustered system right? 16x4 clusters running MPI on linux (the specs are on google somewhere). such a setup requires support for MPI , and the different implementations of this technology on different kernels will yield differing performance levels. im not saying that linux would provide the best results, or that the performance difference between different MPI implementations would be significant - just that it would exist and you would need to choose your operating system based on where you think you'll get the most gains.

the other advantage of open source kernels is that the kernel can be stripped of uneccesary crap, and since hydra runs a chess engine only, this is ideal for maximising memory usage etc. i don't think you can do that (at least to the same extent) on any proprietary OS

i will concede that the use of 'linux' in the posts subject was unneccesary and excessive since it's not the primary governing factor when evaluating such a system, but i don't think the OS is 'meaningless'.

scared

[mnemonic_]

Actually, the time to get scared is when a chess computer becomes sentient, creates an army of robots and enslaves the organic world. Our only hope then will be the chess grandmasters, academic athletes turned heroes of mankind.

Beating a supercomputer is easy..

[Thomas DM]

I bet I can beat every supercomputer on Earth.. If you just allow me to pull the plug ;)

Re:Beating a supercomputer is easy..

[John Seminal]

I bet I can beat every supercomputer on Earth.. If you just allow me to pull the plug ;)

Or a really powerfull magnet.

But then again, I could put some CN in anyones food and have the same effect.

The differance between a person and a computer is people can learn. A computer can not. I played chess for many years, and I did not get better by reading books or studying past games. I got better by playing.

Chess can never be reduced to a number of possible moves just like art can never be reduced to a number of strokes. God gave us something which seperates us from all other things on earth. We are unlike anything else.

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.

Re:Beating a supercomputer is easy..

[shobadobs]

God gave us something which seperates us from all other things on earth.

A bad sense of spelling?

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.

Okay, you just don't know what you're talking about. The whole "unsolvalble geometric figure" thing doesn't exactly work, unless you've got a buggy program. Neither does solitaire. Giving a "sleep" command does seem to work for most computers, though, especially for Microsoft operating systems.

Re:Beating a supercomputer is easy..

[RRRussian]

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.

Re:Beating a supercomputer is easy..

[KDR_11k]

Created by a creator or not, man isn't some kind of super creature that is completely separate from all other life on Earth. Sure, the creator would have created man but he would also have created apes, rocks and air.

Re:Beating a supercomputer is easy..

[drsquare]

The differance between a person and a computer is people can learn. A computer can not

Computers are getting better and better. AI is advancing constantly. Humans aren't getting any cleverer. Give it a hundred years, and we'll be obsolete.

Re:Beating a supercomputer is easy..

[incom]

"...I wonder if anyone has found a way to force a shutdown loop, to do something so illogical the computer can not continue."

I've done very well at large chess tournaments in the past by performing really wacky, yet safe, openings, works especially well against the type that like to approach the game like a science by reading books and memorizing openings etc. The less conscious parsing of chess, and the more intuitive parsing, the more magical your skills become! Sucks I haven't played in so many years, all the chess engines available for linux can kick my ass at the hardest settings, but I'm hoping with some practice I can regains my skills.

Re:Beating a supercomputer is easy..

[RzUpAnmsCwrds]

"Chess can never be reduced to a number of possible moves just like art can never be reduced to a number of strokes. God gave us something which seperates us from all other things on earth. We are unlike anything else."

Wrong wrong wrongedy wrong. We are very much like many other things on Earth. You only need to look at the physiology of an ape to see that. Of course, we are not apes. Indeed, we are unlike everything else - but, indeed, so are apes. It's like claiming that my PC is unique - indeed, my custom-built system is "unlike anything else". It's meerely a property of a specific classification. You, on the other hand, are trying to make it into something that it is not.

"Chess can never be reduced to a number of possible moves"

Recent competition seems to have disproved that. Indeed, it is possible to be highly effective at chess merely by evaluating vast numbers of moves.

Does that tell us that we are less smart? Does it indicate that computers are "superior" to humans? No. It merely indicates that chess is not the ultimate test of thinking that we thought it was.

Someday, however, we may have computers that can paint. Computers that can "think". Do not believe that there is some "divine difference" that makes us the only things capable of thought. There isn't.

Re:Beating a supercomputer is easy..

[Bastian]

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.

This isn't Star Trek. What you suggest is impossible because the chess computer is not trying to guess what the other person is doing or interpret the moves on the board in any other way. It is simply solving a heuristic function based on the positions of the pieces on the board, and the output of that function is the computer's next move.

Re:Beating a supercomputer is easy..

[Mornelithe]

Science is only a bunch of theories.

A theory is a logical explanation of observable facts that serves to explain the world around us. Scientists pick the theories that best explain the facts we've observed so far, and when facts that contradict those theories arise, they will refine them or perhaps come up with all new theories that will once again adequately explain our observations.

What is your process?

We are not apes. We are humans.

From what we've observed so far, the brain is made up of a lot of very small units that take inputs, perform various computations, and fire outputs based on those computations. When you link billions of them together in very complex ways (and allow them to develop new links over time) you get human thought and learning.

What makes you think that an ape's brain is made up of significantly different material?

Today, our computers aren't powerful enough to simulate billions of these units with massively complex interactions. However, it isn't inconceivable that computers in the future could simulate such massive, complex systems, with all the relevant growth conditions. Therefore it isn't inconceivable that we could, in the future, build a computer that learns (provided we learn what exactly goes on within the neurons, and how they fit together).

What are your arguments against this?

Re:Beating a supercomputer is easy..

[Dun Malg]

I've done very well at large chess tournaments in the past by performing really wacky, yet safe, openings, works especially well against the type that like to approach the game like a science by reading books and memorizing openings etc. The less conscious parsing of chess, and the more intuitive parsing, the more magical your skills become!

Much like the way unskilled "button mashers" can initially beat seasoned players of Street Fighter type games with surprising ease. It is, unfortunately, not an effective long-term strategy.

Re:Beating a supercomputer is easy..

[doubledoh]

Now that is comedy.

Seems like man is mauling machine ...

[DrJimbo]

... as the web site crawls to a halt.

Yeah, but...

[chriswaclawik]

In my mind, there will always be only one true grandmaster.

I'm still waiting for the day where a supercomputer can win a rap battle against a human...

Re:Yeah, but...

[AndroidCat]

Maybe not, but the current champs of Dance Dance Revolution should feel threatened.

Re:Yeah, but...

[FidelCatsro]

Yo , yo is talkin trip Holmes , Terminator X is da Big Blue to his Enigma machine -
X gonna go cold lampin on yo an yours , and turn yo kasparov

I agree, i'd love to see a computer defeat a 9dan

[Anonymous Coward]

go player..

Also a lil info from Wiki on Go and Computers:

Computers and Go

Main article: Computer Go

Although attempts have been made to program computers to play Go, success in that area has been moderate at best - development in this area has not reached the level of Chess programs. Even the strongest programs are no better than an average club player, and would easily be beaten by a strong player even getting a nine-stone handicap. This is attributed to many qualities of the game, including the "optimising" nature of the victory condition, the virtually unlimited placement of each stone, the large board size, the nonlocal nature of the Ko rule, and the high degree of pattern recognition involved. For this reason, many in the field of artificial intelligence consider Go to be a better measure of a computer's capacity for thought than chess.

Use of computer networks to allow humans to meet, discuss games, and play one another, although generally considered inferior to face-to-face play, is becoming much more common. There are servers and software to facilitate this; see Additional Resources below for more information.

Linux Chess Supercomputer Overpowers Grandmaster?

[EvilMonkeySlayer]

Phew, for a minute then I thought the machines had risen and were exterminating mankind.

You see, for some people we don't just not RTFA, but we also don't RTF subject.

Often i'll not even read the title and just imagine up my own interesting news for nerds.

Like:

Monkeys become sentient and megalomaniacal. Invades Sweden for no apparent reason.
RIAA sues *insert file sharing company here* the *insert organisation name* is outraged, *insert frail child or elderly person* shocked.

Overpowering?

[nastro]

After overpowering the Grandmaster, the supercomputer chess-bot was heard to say

"I am Bender, please insert girder"

Computer vs. Computer

[alewar]

It can be interesting to see a championship of computers vs. computers, with similar technology but different programming.
To watch a computer defeating a man playing chess is not even interesting anymore, is like trying to do multiplications faster than a calculator (I know some people claim that).

Re:Computer vs. Computer

[Eric119]

Actually, there have been computer vs. computer competitions for many many years.

impressive

[Antonymous Flower]

now if we can just get a machine to 'confirm you're not a script' by typing the seven letters shown in the image.

A far better contest is compression.

[Baldrson]

Compression is a far better basis for intelligence competition than chess, the Turing test or even SAT verbal analogy tests.

Marcus Hutter's AIXI paper provides a proof that if an agent is a good model for human behavior, and the universe is computable, that the most intelligent program is the smallest program that losslessly compresses the set of observations of the universe.

I've formalized a prize competition based on this criterion as the C-Prize, modeled after the Methusela Mouse Prize. The big difference is that instead of lifespan the metric is intelligence. Here is the currently published C-Prize criteria:

Since all technology prize awards are geared toward solving crucial problems, the most crucial technology prize award of them all would be one that solves the rest of them:

The C-Prize -- A prize that solves the artificial intelligence problem.

The C-Prize award criterion is as follows:

Let anyone submit a program that produces, with no inputs, one of the major natural language corpora as output.

S = size of uncompressed corpus
P = size of program outputting the uncompressed corpus
R = S/P (the compression ratio).

Award monies in a manner similar to the M-Prize:

Previous record ratio: R0
New record ratio: R1=R0+X
Fund contains: $Z at noon GMT on day of new record
Winner receives: $Z * (X/(R0+X))

Compression program and decompression program are made open source.

Explanation A very severe meta-problem with artificial intelligence is the question of how one can define the quality of an artificial intelligence.

Fortunately there is an objective technique for ranking the quality of artificial intelligence:

Kolmogorov Complexity

Kolmogorov Complexity is a mathematically precise formulation of Ockham's Razor, which basically just says "Don't over-simplify or over-complicate things." More formally, the Kolmogorov Complexity of a given bit string is the minimum size of a Turing machine program required to output, with no inputs, the given bit string.

Any set of programs which purport to be the standards of artificial intelligence can be compared by simply comparing their Artificial Intelligence Quality. Their AIQs can be precisely measured as follows:

Take an arbitrarily large corpus of writings sampled from the world wide web. This corpus will establish the equivalent of an IQ test. Give the AIs the task of compressing this corpus into the smallest representation. This representation must be a program that, taking no outside inputs, produces the exact sample it compressed. The AIQ of an AI is simply the ratio of the size of the uncompressed writings to the size of the program that, when executed, produces the uncompressed writings.

In other words, the AIQ is the compression ratio achieved by the AI on the AIQ test.

The reason this works as an AI quality test is that compression requires predictive modeling. If you can predict what someone is going to say, you have modeled their mental processes and by inference have a superset of their mental faculties.

Mechanics The C-Prize is to be modeled after the Methusela Mouse Prize or M-Prize where people make pledges of money to the prize fund. If you would like to help with the set up and/or administration of this prize award similar to the M-Prize let me know by email.

Re:A far better contest is compression.

[cookie_cutter]

The AIQ of an AI is simply the ratio of the size of the uncompressed writings to the size of the program that, when executed, produces the uncompressed writings.

First off: cool prize idea, I find it quite interesting.

But, should the measure of AI be the ability to losslessly compress data? In many (probably most) real world situations, lossy compression is not only acceptable, but sensible, since exact compression is often more effort than it is worth. Perhaps you should parameterize the problem such that you take into account the penalty for not reproducing the input exactly.

Re:A far better contest is compression.

[pthisis]

Marcus Hutter's AIXI paper provides a proof that if an agent is a good model for human behavior, and the universe is computable

You can stop right there; as far as modern science can tell, the last assumption is invalid.

Re:A far better contest is compression.

[nacturation]

This representation must be a program that, taking no outside inputs, produces the exact sample it compressed.

And this defined intelligence? What a load of shit! I can't even reproduce the exact sample from an arbitrarily large corpus of writings. Are you telling me that your contest will declare gzip more intelligent than I?

Re:A far better contest is compression.

[Anonymous Coward]

"If you can predict what someone is going to say, you have modeled their mental processes and by inference have a superset of their mental faculties."

If only one sample is compressed and decompressed, where's the predictive requirement of that? I'm not sure that superior compression could not be achieved purely mathematically, by statistical analysis, rather than by "predictive" means.

If you gave a first sample, a compression/decompression program was produced from that, then you gave a 2nd, entirely different sample in the same natural language, wouldn't the compression ratio achieved for that using the same program be a much better indicator of real insight?

Re:A far better contest is compression.

[Baldrson]

Practically speaking this is an assumption that any computing application makes about the universe it modeling.

Re:A far better contest is compression.

[Baldrson]

Are you telling me that your contest will declare gzip more intelligent than I?

Of course not. Ignoring fatigue, you can do the gzip algorithm trivially in the same order of time as a computer -- its just a scaling constant. What the machine cannot do trivially is derive the higher order constructs of natural language that humans find it relatively easy to derive. AI's are now getting close to deriving rules of grammar but they're still not there yet. There are higher order rules yet -- such as knowing that when someone classified as a "libertarian" is saying something, certain ideas are more likely to be expressed than when somenoe classified as a "communist" is saying something.... and so on.

Now, how one would set up such a competition to show humans are superior to a particular AI, as opposed to comparing AI's is a good question. If you view the competition as coming up with descriptive rules (sort of the way Cyc does with humans) of conceptual schemata then it makes sense to let humans write programs that do compression and let programs write programs that do compression. The differences in the predictive power of the models, represented by the programs, will drive the compression ratios.

Re:A far better contest is compression.

[Baldrson]

One of the chief tasks of total quality management is the characterization of error. That's one reason why lossless compression is more valuable than lossy compression.

More importantly though is the fact that as soon as we decide to allow lossy compression we have no operational metric as to which AI is superior.

AI has been drifting without a good metric. Hutter has provided it. Actually, if they'd listened to Ockham they'd have had it but they were too busy trying to get government grants to bother thinking about what a 13th century Friar told them to do.

Re:A far better contest is compression.

[pthisis]

Practically speaking this is an assumption that any computing application makes about the universe it modeling

No, it isn't, and many attempted AI implementations don't make that assumption.

OS used is irrelevant

[anything lemon]

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.

Re:OS used is irrelevant

[dvdeug]

Regardless of which operating system was used in this chess match, the sole determining factor is the hardware.

You should read Feng-hsiung Hsu's "Behind Deep Blue", by the author of much of the code. He goes into great detail about why the software mattered, a lot.

Remember that Deep Blue defeated Kasparov with the more aesthetic MacOS

Deep Blue wasn't running MacOS; it was running on a RS/6000 running AIX. Read the above book, or check out http://en.wikipedia.org/wiki/Deep_Blue.

Linux zealots will cling to this "small victory", but software is only a means to an end.

If the means was needed, I fail to see why it isn't important. Almost everything is just a means to an end; does that mean that building working airplanes is unimportant since it's just a means to an end?

If Linux was the considered the best operating system to run the chess program on, I fail to see why Linux programmers shouldn't be proud of that.

The computer did it?

[glwtta]

Here's the thing, all the computer "did" was run electricity through a circuit - an electric heater does the same thing.

The hardware and software engineers who built and programmed that computer were the ones who achieved the victory - the computer has no understanding of chess, nor in fact any capacity of understanding.

Now if they designed a general purpose AI that then learned to play chess and trounced a great-grand master (or whatever they are called), that would be a computer defeating a human.

Re:The computer did it?

[btgreat]

I have a feeling that a computer that learned to play chess from general purpose AI would become self-aware and dominate a healthy share of the world far before it was able to defeat a chess grandmaster.

I still want to know why big companies are spending money building computers to play chess when they could, in theory, be spending money building computers to take over the world. That isn't against international law as far as I know.

Re:The computer did it?

[glwtta]

Well yes, being self-aware was kind of a requirement for the AI I was talking about.

And no, even in theory they don't have the slightest clue about how to build such machines.

Re:The computer did it?

[Anonymous Coward]

This explanation, the "No matter what it does, I hereby claim it's just a big calculator" type of explanation leaves a number of puzzling problems for its proponents.

1. If the capability is all in the engineers, why don't they beat grandmasters ? (FWIW they don't, but you can go do your own digging if you want to reassure yourself that this is the case). The machine is even capable of correctly playing a (winning) ending where real grandmasters OR engineers would agree a draw. They can't see the way out but the machine can.

2. If I teach a child to play chess very well, is it really -me- who is defeating the child's opponents?

3. Why are you so sure that you are not just an expensive chemical heater which happens to peform calculations fixed by someone else? Just because the illusion is convincing to you doesn't mean anyone else should believe you.

Hofstadter was surprised that we proved able to produce a "dumb" computer that beats grandmasters at chess, but perhaps he shouldn't have been. So far as we can tell the grandmaster is also just a "dumb" computer, albeit a fantastically complicated one, and chess is a pretty specialised sub-function of that grandmaster. So it makes sense that a relatively uncomplicated "dumb" computer can match this one specialised task.

BTW you may wrongly imagine that you're a much more subtle beast than any computer, and that your behaviour is very hard to predict. This isn't true, nearly all human behaviour is boringly predictable, but the person doing the behaving never notices! This suggests that our experience of consciousness is at least partly a lie, we do many things without any control over them and they are retrospectively explained away by the conscious mind.

Re:The computer did it?

[Tyler Durden]

1. If the capability is all in the engineers, why don't they beat grandmasters ? (FWIW they don't, but you can go do your own digging if you want to reassure yourself that this is the case). The machine is even capable of correctly playing a (winning) ending where real grandmasters OR engineers would agree a draw. They can't see the way out but the machine can.

No because they can't memorize and evaluate their created algorithms like the computer can. The computer is perfect at the mindless aspects of playing chess. The important part, the creation of the algorithm, is completely in control of the human engineers.

2. If I teach a child to play chess very well, is it really -me- who is defeating the child's opponents?

Because you simply start off teaching the child basic strategies while the child figures out the rest himself. In contrast, the computer's actions are dictated completely by the algorithm.

3. Why are you so sure that you are not just an expensive chemical heater which happens to peform calculations fixed by someone else? Just because the illusion is convincing to you doesn't mean anyone else should believe you.

Because we possess consciousness. It is meaningless to say that the experience of consciousness is a "lie". You need to assume consciousness exists in the first place for there to be something to be lied to.

Re:The computer did it?

[gargletheape]

Neither here nor there. Unless it is your claim that DNA machines which ingest oreos and reproduce can be conscious but silicon chips encased in a cuboidal box cannot.

In any case, it is far from clear that the phenomenon of consciousness that allows a Kasparov to do the cogito ergo sum thing is what makes him a superb chessplayer.

Re:The computer did it?

[Tyler Durden]

What are you talking about?

I believe that a series of silicon chips can become conscious. However, I don't think consciousness can ever be completely defined by an algorithm. Probably a complex enough computer can be given a learning algorithm and then by interacting with the environment it becomes conscious on its own.

The whole argument is about who is really the victor - the engineers who created the algorithm or the computer that performs it. The original poster was arguing that it is the engineers, and because the engineers were the ones with consciousness I believe he is right.

Re:The computer did it?

[gargletheape]

1. How does the complex algorithmic machine become conscious (therefore more than "just" algorithmic) simply by "interacting with the environment", and in what *non-algorithmic* terms can you characterize this interaction? Do you believe that we'll have to find new, non-algorithmic laws of nature simply to resolve the problem of consciousness?

2. About the engineers being the victors because they alone possess consciousness - what if the designers simply designed some sort of genetic algorithm where the computer in fact figured out a near optimal algorithm on its own? What if they had just set up a net of some sort and trained it? Does 'who won' in this instance depend upon what precise method the programmers used?

Re:The computer did it?

[Tyler Durden]

1. How does the complex algorithmic machine become conscious (therefore more than "just" algorithmic) simply by "interacting with the environment",

In the same manner that the combined biological machinery of cells began to manifest consciousness by interaction with the environment through evolution. A lot of people seem to think that if you just get together enough speed, circuitry and memory together the computer will automatically get a consciousness. I think it would more likely need to be forced to adapt to a chaotic environment as well for this to happen.

and in what *non-algorithmic* terms can you characterize this interaction?

I don't know what you mean by this. If it were possible to characterize an interaction enough you'd have an algorithm. I don't think this can be done in this case.

Do you believe that we'll have to find new, non-algorithmic laws of nature simply to resolve the problem of consciousness?

I don't think consciousness can be fully modeled in terms of neatly-defined laws of nature. I'm not saying that consciousness is supernatural, just that any attempt at a rigorous description misses the point. It is impossible for mechanical descriptions to explain experience.

The best resolution I think there would be to the problem of consciousness would be, "Given an algorithm with properties a, b, c and such-and-such tasks for it to perform in so-and-so environment, consciousness is likely to occur.

2. About the engineers being the victors because they alone possess consciousness - what if the designers simply designed some sort of genetic algorithm where the computer in fact figured out a near optimal algorithm on its own? What if they had just set up a net of some sort and trained it? Does 'who won' in this instance depend upon what precise method the programmers used?

Hmmmmmmmmm. A fiendishly good question as this concerns a question of degree. In both your examples I would say we are moving more towards the computer winning than the engineers. (Which is probably what the original poster was getting at). So it would be a continuum with the computer following a database of moves at one extreme to a conscious computer figuring out chess strategy at the other. I would definitely say that the methods the programmers used matters.

My brain hurts.

Re:The computer did it?

[ahdeoz]

Wrong. People like you are boringly predictable, but many others are fascinating and unpredictable, capable of making both arbitrary and reasoned decisions. Don't project your stupidity and lack of consciousness on the rest of us.

Re:The computer did it?

[glwtta]

3. Why are you so sure that you are not just an expensive chemical heater which happens to peform calculations fixed by someone else? Just because the illusion is convincing to you doesn't mean anyone else should believe you.

I never claimed any such thing, in fact I am absolutely positive that I am just such a chemical heater (well, I would take issue with being programmed by someone).

The point is that we know that the engineers provided the fixed algorithms for the chess computer - in this instance it's still the human heaters doing the work, regardless of how they themselves were programmed.

And please, you make far too many assumptions about what I would assume. (btw, I didn't bother with 1 and 2, I think it's pretty obvious why they are irrelevant)

Re:The computer did it?

[julesh]

Neither here nor there. Unless it is your claim that DNA machines which ingest oreos and reproduce can be conscious but silicon chips encased in a cuboidal box cannot.

That's not necessary: all that is necessary is to claim that this particular computer is not. And as the algorithms it uses to play chess are all known and understood by its engineers, it is likely they would know if it was conscious, and that being such a great achievement, would have announced it by now. ;)

Re:The computer did it?

[caranha]

The worst part about AI research is that, whenever we manage to do something people consider to be AI, they say that this thing is not "real" AI, that something else a little ahead is "true" AI.

The above (paraphrased) isn't mine, but I agree fully.

Sigh :-)

Re:The computer did it?

[glwtta]

That just highlights the problem I was talking about - what's termed 'AI' in research, is not what we think of as 'AI' intuitively. That's because there isn't any research into actual AI (ie artificially creating what we understand to be sentient intelligence), rather all that present AI research does is approximate some aspect of the appearance of intelligence.

I would be really surprised if we had something which was only a little behind "true" AI.

I think the real problem is that we (ie programmers and related proffessions) are so attached to thinking about our computers in anthropomorphic terms that we throw about the term 'AI' far too quickly.

Re:The computer did it?

[Ohrion]

Hmm... I still think that even with a general purpose AI the victory would still go to the programmer. Maybe it should go to the AI Trainer as well. SOMEone has to walk the AI through how to do basic things.

Yeah, so?

[ian rogers]

I could beat the computer in a boxing match.

Anyways, everybody knows a pound of muscle weighs more than a pound of brains.

Re:Yeah, so?

[Bob 535604]

everybody knows a pound of muscle weighs more than a pound of brains.

What? They both weigh a pound! How can muscle be heavier?

Re:Yeah, so?

[Chandon Seldon]

you = pwned

New movie about it coming out?

[AndroidCat]

March of the Penguins

All your king are belong to us

[Avalonia]

How does the human rate on performance/Watt compared to the machine? Isn't that what's important these days?

Re:All your king are belong to us

[zippthorne]

Moderately hungry humans use about 100 watts, on average, throughout the day.

Yawn...

[flood6]

So what. Chess is so one-diminsional. I wonder how good that machine would do at a real skill game, like Rock, Paper, Scissors.

Re:Yawn...

[BlueNeutrino]

So what. Chess is so one-diminsional. I wonder how good that machine would do at a real skill game, like Rock, Paper, Scissors.

Even the slowest machines will easily defeat any human (humans can't be truly random, while machines can do that a lot better)

Re:Yawn...

[_mt99]

Have a go at Roshambo.

Re:Yawn...

[slashkitty]

ah, but does truly random win in this game? I bet a human can beat a human better than a computer can.

Re:Yawn...

[BlueNeutrino]

Well, being truly random helps you not *losing* this game. The moment you find that your opponent has somehow outsmarted you, you can simply switch to the "random" mode. On the other side, a computer can easily model its opponent as a markov source and learn some statistics. You'd be suprized how predictable humans turn out to be... The computer can beat a human "better" because it learns more accuratly (because of the better and more objective memory). There are a some games that computers truly suck at (such as Go or Poker), but Rock, Paper & Scissors isn't one of them.

Re:Yawn...

[AntibubbleAli]

I am not saying RPS is not a real skill game. I agree it is a more social game than a computer could probably pull off and play well, but it could be programmed to keep track of patterns a person could subconsciously be doing, and apply a RPS gambit (http://www.worldrps.com/gambits.html) to try to win. I've tested these strategies against a computer that was picking random choices and some worked better than others. So, if a person was really choosing by random, maybe some of the strategies would work if a machine used them. But, since our brains are always picking out tiny patterns and can not do something 100% randomly, could a program potentially pick up on subtle patterns? I never got a chance to have Rock, Paper, Scissors battles between basic programs to play the games, but I am also curious how they'd do against each other or if more complex programs could do reasonably well against a human...

Re:Yawn...

[SnprBoB86]

You, my friend, must be a member of the World Rock Paper Scissors Society.

http://www.worldrps.com/

You've got to try the "Online trainer." Read all about it's sophistication before playing.

Best of all IT'S OPEN SOURCE!

An increadable piece of AI technology completely written in Javascript. Check out the page source when you are done playing.

Re:Yawn...

[SnprBoB86]

Aw, I'm a dumb ass. I remembered incorrectly. The source is server side :-(

but you can clearly get the joke from the a file name in the page source...

Re:Yawn...

[wombat_of_doom]

Now there's something I never would have thought existed. Strategies for playing Rock Paper Scissors! Could be quite interesting to have a computer program play, especially if it is designed to learn and improve against the player. Let us know when you've got it written. ;)

Big Fucking Deal

[suzerain]

Chess is a game of pattern matching that requires basically no abstract though (well, OK it requires a little, but for the most part, it's just data recall). Computers ought to trounce human beings at this sort of endeavor.

Wake me up when a computer can beat a Go master...

Re:Big Fucking Deal

[Dpaladin]

The best human chess player is still better than the best computer. I guess the relevance of this chess machine is that it wasn't designed specifically to beat one opponent, like Deep Blue was.

Re:Chess is only for humans

[blue trane]

You are always free not to play against a computer.

Basically this type of accomplishment means that another task we used to consider indicative of intelligence is not really as important to intelligence as we once thought.

We're good at learning new games, learning rules, adapting strategies. When we can get a computer to do that, that's when you start worrying...

Deeply ashamed

[MyLongNickName]

I am deeply ashamed that we have developed machines that are stronger than us. Clearly, because machines can outlift and outwork us, we have lost our purpose as humans.

And computers can outcalculate us in highly linear situations. It is time to pull the plug on humanity, and let the chess programs and heavy lifting equipment to collaborate...

Re:Deeply ashamed

[btgreat]

If this is true, then what is the purpose of monkeys?

Think of where we will be once the machines take over. I don't think it will be like in the matrix where we tried to fight back, but rather like nature. Humans will begin to show their inferiority, and will run rampant swinging from trees and eating bananas. The machines will open zoos of humans, and do product testing on us.

I, for one, am planning on being a pet. I'll take the loving grasp of my artificially intelligent machinistic overlord over a banana and loincloth any day.

Top 100 list??

[John Seminal]

The first american is not listed until #19. That is not possible. There are 7 russians on the list before the first american. This has to be a lie.

Let me tell you my logic.

The USA has the biggest economy, the best army, we do everything the best. It is not like we steal or lie or cheat to deprive others of what is theirs.

Okay... there was a heavy element of sarcasm there.

But honestly, looking at that list, is there anything it can tell us about a countries intellectual power? Or could it be just as easy to replace the title of the paper from "Top 100 Chess players" to "Autism list"? Does being good at chess correlate to anything else? Does chess score indicate IQ? Does chess scores indicate earning power? Anyone have a t-shirt that says "I play chess... Love me before I become rich"?

And why is Bobby Fisher not listed. Do people lose their rankings if they don't play for a set time. It seems only fair for a person to keep their ranking when they retire.

And this gives me one interesting idea. I wonder if instead of a super computer, we can recreate one man. If we can take the thinking process of one person, and program a computer. That way, people who live 400 years from now will be able to play me.

Re:Top 100 list??

[nomadic]

Does being good at chess correlate to anything else? Does chess score indicate IQ? Does chess scores indicate earning power? Anyone have a t-shirt that says "I play chess... Love me before I become rich"?

There is a definite correlation between skill at chess and interest in playing chess. That's pretty much it.

Re:Top 100 list??

[fishbowl]

"There are 7 russians on the list before the first american."

A single generation ago, chess was a national priority of the Soviet Union. It was routine for young people to be evaluated on their chess aptitude, and if they demonstrated it well, they'd go to chess camps. If they performed among the top ranks of the most serious players here, they'd be groomed as potential champions.

In the USA, chess has never been considered as more than a game, certainly not to the point that there is a national priority placed on developing chess masters. And American society does not really tolerate the idea of a person being coerced into a particular profession based on his ability, and at the discretion of some government bureaucrat.

The decline of the soviet system may mean a more level distribution of chess granmasters in future generations. Or it may mean that the very process by which grandmasters are identified early enough to get that kind of training, is gone.

I'm going to ignore your more cynical statements about the USA.

Re:Top 100 list??

[blue trane]

Maybe they can train them to write programs to play Go now...

Why a Computer Can't Win. (Usually)

[Dixie Flatliner]

I rarely post, but I thought this might be worth reminding people of.

While computers are easily tactical masters of chess playing - in that they can immediately anaylze all possible moves availible in a given play, and determine possible outcomes, their fallacy comes in strategy, because, put simply, they don't know how to win.

What is a good move? Is it one that results in a opposing piece's defeat? If so - what value should that piece be assigned? Indeed, what is the value of _any_ piece at any given time on the boards - why should a machine choose one set of perfect moves over another - in almost every way a computer cannot determine the long term value of a move.

This is remedied somewhat by having pre-played game analysis at the disposal of the machine, but in almost every case the computer program requires serious recalibration between matches to prevent a human player from adapting to a strong tactical game. It is by no stretch that computers can be considered inferior in almost every way to a strong human player.

Kasparov posited Advanced Chess as the ultimate play form; the tactical mastery of a computer, mixed with the multilevel strategy of a grandmaster player, making for a game of sublime subtley and perfection.

Re:A game where computers will never exceed humans

[ettlz]

Apu Nahasapeemapetilon did write a program for his Ph.D. that could beat all but the toughest grandmasters at Tic-Tac-Toe (or, Noughts and Crosses, to give it its proper name). Unfortunately, it was written on several hundred punch-cards, and lost to an accident involving Bartholomew J. Simpson.

nah you can just stay asleep

[layingMantis]

And when (not if), GO AI has progressed to the point of beating really good humans, you'll yawn and say wake me up when they can play So what's your fucking point?

I think we'll just let you sleep :)

~mantis

Now have it "outsmart" Kasparov

[melted]

Then we'll see. Until then you can't even reliably say that computers are better than humans at playing chess.

Hydra is not the ultimate chess entity

[Redshift]

Hydra performed very badly in the recent Freestyle Chess competition run by ChessBase - competitors were a variety of grandmasters and amateurs assisted by databases and computers. I other words, any form of cheating was acceptable, all that mattered were the moves on the board.

The two Hydra machines did not even make it into the final sixteen. Moreover, the eventual winners were a couple of amateurs using pretty ordinary PCs running over-the-counter chess programs. On the way to the title they beat a selection of computer- and supergrandmaster-assisted grandmasters.

On this evidence the "strongest chess entity on the planet" is a team consisting of a New Hampshire database administrator + a soccer coach + 3 ordinary PCs.

Links:

Hydra knocked out

Final result

Winners debriefing

Re:Hydra is not the ultimate chess entity

[Bastian]

Agreed. I don't put much stock in these much-touted computer chess programs beating Kasparov and the like, because the computers are allowed to effectively cheat by having massive databases of chess moves on file, and using them they can simply copy the strategy and tactics thought up by other humans. And that is how they win.

I imagine that if you took Hydra or Deep Blue's game databases away, they would in fact perform very poorly.

Re:Hydra is not the ultimate chess entity

[mesterha]

I don't put much stock in these much-touted computer chess programs beating Kasparov and the like, because the computers are allowed to effectively cheat by having massive databases of chess moves on file, and using them they can simply copy the strategy and tactics thought up by other humans. And that is how they win.

So are humans cheating by using the ideas thought up by other humans. Chess programs are programmed, so it doesn't make sense to assume experts aren't allowed to help and that databases can't be used.

The complaint Kasparov had was that the program wasn't fixed. In other words, programmers would change the program in between matches. This is clearly unfair in the sense that he is really taking on a computer/human hybrid if the program is allowed to be modified during the tournament.

Re:Hydra is not the ultimate chess entity

[Luthair]

Amateurs assisted by over-the-counter chess programs probably make unexpected moves that Grandmasters and Supercomputers aren't experienced with countering.

Re:Hydra is not the ultimate chess entity

[damiam]

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.

Re:Hydra is not the ultimate chess entity

[Rick and Roll]

I agree about the computer/human hybrid remark.

I think a researcher should find how much information about chess moves a human is capable of storing and limit the in-game databases to that.

As for experience, I think on top of a limited amount of recallable knowledge, the programmers should be able to throw lots of data at the computer that doesn't wind up being accessible in its original form, but is used to train neural nets.

This would make such a competition more interesting for me.

Re:Hydra is not the ultimate chess entity

[Bastian]

I think that's been done because all attempts to use neural nets in chess computers (that I know of, anyway) have not been particularly successful.

But then, possibly this is because there hasn't been as much effort put into neural approaches.

game 37 on fics

[sfcat]

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.

Re:Chess is only for humans

[MasamuneXGP]

Chess stopped being for humans a long, long time ago. Ditch chess and start playing Go. It's true that computers will probably master that as well, but that day is still quite a ways off. The history of Go is filled with great achievements...

Why is this news

[Sigius]

And why does it matter the thing runs on Linux ?

a humiliation for mankind ?

[atari2600]

Get over it. Humiliation for manking is killing each other in the name of religion, oil and what not. Goddamn.

what IS a grandmaster?

[ASLayerAODsk]

so...heres a question, whats the point to ranking this stuff? 'grandmaster' 'super grandmaster' ya know...it sounds like marketing to me more than anything...as is most things these days...how about 'some boob who doesnt get laid much and plays more chess than with himself'? id watch a match with a title like that...even for a laugh :D LONG LIVE HOWARD STERN!!!

obFuturama Quote

[Greyfox]

Conan O'Brien's Head: Yeah, well at least I have something you'll never have! A soul!
Bender: Big deal!
Conan O'Brien's Head: And freckles!
Bender: (crying) Whaa...ha..ha...

Fisher Random

[m0rphm0nkey]

Bobby Fisher may be a bitter anti setmitic inhabitant of the lunatic fringer, but he's right about one thing. The reason computers beat humans isn't because they think better, it's because they don't forget stuff.

If fisher random or one of the other shuffle chess variants were used, computers would most likely only make mediocre opponents.

Deep blue cost how much? The price of a human chess player is only whatever it costs to raise her/him. Undoubtedly still a much more efficient use of resources, and (having children myself) a much better excercise in patience.

m

Love the word choice...

[writermike]

"Overpowered," huh? I immediately imagined the machine LOSING, then transforming into something like Optimus Prime and overpowering the GrandMaster.

Cue Transformers music bumper.

You may have been joking.

[Anonymous Coward]

But rock paper scissors is a very interesting computer challenge because it Dan Egnor's Iocaine Power

demolished the field of the international computer rock paper scissors contest with its 6 levels of sicillian reasoning. (I know what you're thinking; but you know that I know what your're thinking; but since I know that you know that I know what you're thinking - I can beat you).

Female grandmasters

[5n3ak3rp1mp]

Actually, what I thought was more impressive than this was noting the number of female grandmasters in that FIDE link above. It's about time! (Is the fact that none of them are American have something to say about cultural discouragement of intellectual excellence in women here?)

Re:Chess is only for humans

[ImagistTD]

What's not to like? Sex isn't any worse just because porn sites were invented.

This stuff isn't interesting any more

[gelfling]

Any more than races between horses and horseless carriages. I mean people are not machines, get over it.

Wooden Nickels

[headkase]

Compression is a far better basis for intelligence competition than chess, the Turing test or even SAT verbal analogy tests.
Or view intelligence as a signal to noise ratio with more signal representing a greater degree of some type of information. Compression is not intelligence, it is representation.
This representation must be a program that, taking no outside inputs, produces the exact sample it compressed.
All compressed representations of information as we know it today compress by replacing a percieved order with smaller tokens with the effect that the data becomes more and more random until it has reached it's maximum compression and no new order can be tokenized.
The reason this works as an AI quality test is that compression requires predictive modeling. If you can predict what someone is going to say, you have modeled their mental processes and by inference have a superset of their mental faculties.
Compression is not predictive. While there is a grain of truth with prediction, it's more like intelligence is the ability to predict and apply the predictions to create a desired outcome within your environment. It is dark. I predict flicking that switch will create light. Flick's switch. And to model someone's processes to the detail that they do your would need about three pounds of meat plus your extra process modeler to do so accurately. Of course there would be ethical problems with doing so.

OS used IS relevant for different reasons

[cookie_cutter]

EVERYTHING is a means to an end. The important thing is what kind of software has the most utility. Free open source software has (all other factors being equal) more utility than proprietary software because people are not inhibited from building upon existing work since they aren't encumbered by lack of 'copyright' rights as they are with non-free software.

Re:Can afford his chess library

[notthepainter]

So yes - openning databases are known quite deeply by the best players - a computer using a database is only fair.

I had a friend who played chess competively. He eventually stopped and I asked him why. He couldn't afford it, he said. Huh? Airfare to tournaments? I wondered.

"No," he replied, "I can afford to maintain my library."

Yes, his chess library of opening moves. He couldn't keep it at the level he needed it. That certainly brought a new perspective to the game.

MC Hawking

[PxM]

Does this count?

So What Else is New?

[r1_97]

It's John Henry all over again. Computers can add, divide, calculate, compare etc. faster than us humans. Humans also build the machines to lift and drive steel better than their creators. It's the design and creation which distinguishes from the machines we create.

The real reason why Go is hard

[MAdMaxOr]

While the search space may branch more quickly in Go, relative to chess, this is not the primary source of the difficulty of Go. The primary reason why Go is hard is that the results of any move are not fully apparent until the distant future.

A blunder in chess will typically result in a loss of material or a significant measurable disadvantage within five moves or less, and often on the very next move. A blunder in Go may only become apparent forty moves later. Forty moves is well beyond the limits of current technology.

A possible side effect of this (just conjecture), is that it is also much harder to measure the effects of sacrificial moves in Go.

fpgas

[kunzy]

Other than Deep Blue, HYDRA uses dedicated logic, not just a cluster of standard PCs.

Each of the 64 processors in the cluster includes an FPGA (Field Programmable Gate Arrays) card from XiLinx, which are significantly faster than Pentium or Athlon.


Cool.

This is not what it seems

[Anonymous Coward]

Chess matches between humans and computers are not what they seem- the computer has a huge advantage that amounts to a cheat, specifically, the computer has access to all of its opponent's (the human's) previous games, while the human has no such access to the computer's previous games. Typically, the computer is programmed to play against THAT specific player through analyzing that player's past performances. The human, on the other hand, is not given the computer's games to analyze, they have to show up and play "blind", so to speak. This is a HUGE advantage in favor of the computer. This was the case in the Kasp vs. BB match and I am pretty sure it is still true today. Let the human have access to the computer's previous games, let the human study them and then we will have an even playing field.

no privacy

[Sebastopol]

i like how all their birthdates are posted too.

now if i'm ever stealing kasparov's identity, i have the requisit security answer...

heh heh heh.... doh!

This is a somewhat naive Test/Prize ..

[RedLaggedTeut]

Take an arbitrarily large corpus of writings sampled from the world wide web.

These will contain random mistakes and contradictions and as such are probably not very good at creating an intelligence that can not only compress "the universe", but create new intelligent data as well(without repeating itself too much).

I also wasn't aware that his knowledge of physics and chemistry makes anyone better at picking up girls, i.e. at solving granular problems.

I predict that bzip -9 corpus.txt will be the winner of your contest, or more accurately, followed by strip bunzip2 && bunzip2 corpus.txt.bz2, where the pair of (bunzip2 corpus.txt.bz2) is the program.

Of course the real contest doesn't stop there. Because to compute the full information content, you need take into account the context provided by CPU and OS. You could also try to compress the bunzip2 binary, rename it to 'b', but all these savings will be minor compared to the already high compression rate of bunzip2.

At least that is how I am guessing it will turn out.

Re:This is a somewhat naive Test/Prize ..

[pthisis]

I predict that bzip -9 corpus.txt will be the winner of your contest, or more accurately, followed by strip bunzip2 && bunzip2 corpus.txt.bz2, where the pair of (bunzip2 corpus.txt.bz2) is the program.

I'd think lzma would beat it (along with other general-purpose compressions algorithms that have better compression rations than bzip2).

But I wouldn't be too shocked if a superior special-purpose algorithm could be written that worked better on (say) english-language text but far worse on everything else. And such a thing is likely to be somewhat interesting.

Though I still think any AI that acts (to the average human observer) deterministic, let alone computable, is doomed.

Advanced Chess: Human-Computer Collaboration

[FleaPlus]

Seeing so many posts describing this as a "man versus machine" thing compels me to mention advanced chess, a new form of chess recently proposed by Garry Kasparov. The gist of it is that instead of humans and computers working either alone or against each other, a human player and a computer player team up. Personally, I think competitions like that are great for exploring how humans and computers can achieve a better symbiosis with each other, taking advantage of the strengths of each.

From wikipedia:

Advanced Chess is a relatively new form of chess, first introduced by grandmaster Garry Kasparov, with the objective of a human player and a computer chess program joining forces and competing as a team against other such pairs. Many Advanced Chess proponents have stressed that Advanced Chess has merits in:

* increasing the level of play to heights never before seen in chess;
* producing blunder-free games with the qualities and the beauty of both perfect tactical play and highly meaningful strategic plans;
* giving the viewing audience a remarkable insight into the thought processes of strong human chess players and strong chess computers, and the combination thereof.

Re:Advanced Chess: Human-Computer Collaboration

[TopherC]

I love this idea! I was also thinking that an all-human collaboration might be appropriate against a computer. Historically, has there ever been a chess game where a group of people played for one side or both of a chess match?

In a lot of ways it is interesting to see if and how a group of people can effectively collaborate on hard problems like you find in chess. I also wonder how much of chess strategy is in finding a particular cohesive style that has proven to work, and how much is just move-by-move out-thinksmanship. Would a collaboration weaken itself by shifting strategy or style too often? Or would that turn out to be a strength?

A Pretty hollow "Win"

[Mulletproof]

I guess you can call it a win that you've been able to build a computer that can out calculate the human brain. Of course, then you have to calculate the combined manpower, resources and support it took to accomplish that feat.

Brute force calculations are a computers specialty. That's what they're primarily built to function as. Polygons. Spread sheets. Folding@home. All nothing but tools to perform brute force calculations. like chess. Calculate every permutaion and come up with the most efficient response. There is no real overriding strategy. No personal style of play. No attempt to fathom what your opponent is actually trying to do given his playing style. How many millions did it cost again? For something that can only outcalculate a human? 2/3rds of the time? Sometimes less?

Sure, I guess you can claim a win. But then, I have something that can out calculate me too.

I play Battlefield 2 on it.

Re:A Pretty hollow "Win"

[gargletheape]

And what of the grandmaster's multiple trainers, coaching from youth, expensive schooling and healthcare, hours and hours of careful analysis and carefully researched traps sprung on unsuspecting opponents? This machine is better at chess than Adams is. Period. Unless you want to start claiming that the best chessplayers / sprinters / stockbrokers / physicists are no better at what they do than anyone else because it cost "many millions"

Just the Beginning

[umbrellasd]

I remember when people disclaimed the notion that a computer could even beat a grandmaster at all. Now it has happened, so people say, "Well, it is not really the computer doing it. It is the people programming it."

You are right that the computer is just running algorithms, but what if computer scientists write a bunch of algorithms for cooking? This spice combines with that spice in a pleasant way. This amount of meat requires this many minutes at this temperature, and so on. Now suppose that someone programs a computer to invest wisely in the stock market. This has been done already with success. Not grandmaster success, perhaps, but you see where it goes.

A computer can be extended with greater ease than a human. More memory, more CPUs, more disk, and faster of each is always nice. Even with the power that our computers currently have, we can certainly build algorithms for a great variety of things--all our effort, yes--but then we can load all of those algorithms into the memory space of a single computer and add an algorithm to recognize, "I want to play chess." as the time to apply chess algorithms, and "What would be a good recipe with fish and red peppers?" as the time to apply cooking algorithms.

And so it goes. At some point, with such a wealth of algorithms at the computers disposal and the ability to recognize when to apply which ones, you are getting very close to at least "apparent" sentience, at least in the limited sense that many of us understand it. This is not so unreasonable; it does not require much more than what we can do now. More developer time invested in algorithm development across a broader range of topics, and possibly memory/CPU power, though already, computers are rapidly approaching the best case estimates of our brains capacity (from a biological standpoint of number of neurons, rate of synaptic firing, degree of connectivity, etc.). This is not so different from raising a child. Years of teaching algorithms to interact with the environment and society in a beneficial way. We can even program computers to modify their own algorithms based on external simulus (as our children do).

I do not suggest that there are not obstacles. Certainly there are. But as you pointed out, the single biggest obstacle to making a silicon based sentient is likely not the materials*. It is our currently limited, but rapidly increasing understanding of our own sentient processes, that is the real limitation. In other words, the computer is not the limited one. We are.

The next 50 years should be tremendously interesting.

* And if materials were the issue, we could switch to other materials (already happening), or as our understanding of genetic engineering increases, we can switch to a biological substrate.

Re:Just the Beginning

[mesterha]

That's exactly what they said 50 years ago.

Re:Just the Beginning

[glwtta]

Everything you say makes a lot of sense, except each of your conclusions is a bit of a leap.

Yes, it would be exteremely interesting to see what happens when we build a computer (in our current understanding of what that is) that interacts with its environment and learns from it. Whether this produces "apparent" sentience or "actual" sentience would be quite enlightening.

That's not what we are talking about here though, the chess computer does not interact with its environment, has no capacity for learning, all it is is a bunch of heuristics and database lookups. I'm sure so are we, but in this case what the computer does is strictly defined by humans, not built up by learning.

A computer can be extended with greater ease than a human.

I can't agree with that - I spend most of my day programming computers to do new things, I know plenty of people who learn a lot faster than that.

It is our currently limited, but rapidly increasing understanding of our own sentient processes, that is the real limitation.

Of course, this limitation was my entire point. But our understanding here isn't "limited", it's nonexistent. It's not "rapidly increasing", it's making absolutely no progress whatsoever. Oh sure, we are learning plenty about the hardware (neuron function, synaptic pathways, etc), but the last time I checked, we are not one iota closer to having a theory of how sentience, consciousness, or self awareness works (or in fact, what it is).

This isn't some hubristic "we are oh so special" thing, I'm just saying this is something we know jack about, and are nowhere near being able to replicate artificially.

In other words, the computer is not the limited one. We are.

That's really neither here nor there. We are limited (as is everything), and computers, being our creations, are limited by our ingenuity, in addition to all their other limitations. So what?

Fair contest?

[amightywind]

Is it a humiliation or triumph for mankind that it can build a machine that can defeat itself?

Human's should not feel humiliated at all. Indeed a great player like Kasparov can still embarrass the strongest chess program if he is given a little information on how it is programmed. If humans are expected to play these programs blind, and programmers have full access to the game records of the strongest grandmasters, is it really surprising that the humans lose? Kasparov has shown that with solid understanding of the opposing program and access to its playing record it is still possible for the human to compete. The fact that there has been counterattack on chess computers at all since Kasparov/Deep Blue is a tribute to the human mind!

Whats he big deal?

[l'obscurit]

The strength of a computer program has nothing do to with the operating system!?!? IT is about the algorithm used to implement the chess engine. If there is any advantage to an OS it would simply be that one is FA?STER than another and hence able to comple up with it's best movve within the requisite time limit. Seduction Home Page

GO?

[Wooky_linuxer]

Since you said "at this point", point me a computer that can win a match against a Go Master. Yes, it's a game that can be mathematically explained, only it's orders of magnitude more complex than Chess. Computers suck at it. The calculation capacity of a computer will always be finite, so I predict there will always be a game, mathematically explanable, that Men will won. I'm not taking Hobbits into account here, mind you.

Interesting logic

[alexo]

> And, no, chess has not been PROVEN a draw. But please find me any expert
> who thinks chess is a forced win for white... If there were a forced win,
> several hundred years of chess (in current form) would have discovered it.
> Chess programs would have stumbled on it.

By the same logic, if chess is a a forced draw by black, several hundred years of chess (in current form) would have discovered it. Chess programs would have stumbled on it. Yadda, yadda yadda...

> If you disagree... find me any Master level player or higher that corroberates your view.

Find me any Master level player or higher that has solved chess.

That is, unless you accept "proof by concensus", in which case the Earth was the centre of the universe up until the 16th century and the speed of light was not constant until the 19th.

Re:Interesting logic

[gordo3000]

if you mean proof by consensus of the educated class, then the earth has been round well before the 16th century. just to let you know. When columbus sailed the educated class knew quite well the earth was round(at least had more than enough evidence for this) and most sailors believed the earth to be round from experience.

I don't think you got the joke.

[zippthorne]

But i'm not going to explain it.

Sort of expected...

[Eric Damron]

In a man vs machine contest I would expect the machine to win. It's inevitable that as computers become more powerful humans will lose any games played against them. We shouldn't be upset over this. I'll bet I could design a hydraulic jack that could defeat the worlds strongest weight lifter too.

man vs. machine?

[safekd]

I think this is man vs. man. The comuter was programmed with chess technique and theory and game histories, that where developed by man. When a computer wins a grand master with pure calculations , that will mean something.

Computers creating art

[typical]

Art is a form of nonverbal idea and emotion communication. It is probably *very difficult* to build a machine that can produce good art, since the production of art that a human would consider good is simply a problem of understanding the human mind. When I want to see what feelings or ideas a painting that I'm creating evokes, I always have an ongoing human test subject -- me. Computers do not have this.

Creating art from a computer requires not just learning and processing capabilities similar to those of a human but the experience and similar emotions to those of a human, or accurate understanding of such mechanisms.

However, while such a problem is one of the final problems that AI would be able to solve in terms of simulating a human accurately (since the machine have a very accurate model of much of the human mind), there is certainly no reason to think that it is impossible; I see no insumountable barrier.

Re:why this is considered amazing

[fitten]

Why? And why does it matter what OS the thing was running on? Is there something magical about the code to where it can't be compiled and work on any other OS that exists? or is it that Linux somehow contains some magic that makes programs that play chess run better than any other OS that exists?

Re:How do they calculate the players score?

[antispam_ben]

The player with the most kings wins. The score is either 1 to 0 or 0 to 1. Unless it's a tie.

In sportscaster terms:

"In hot chess action today, Deep Blue pummeled Kasparov."

Game 1

[spamguy]

I looked at the results of Game 1 -- game end in 33 moves. 'Awesome,' I said. But then I plugged the results into ExaChess, and was very confused. The game just...stopped, and just when it was getting interesting. If the computer won through time (and I'm not sure there was a timer at all), Mr Adams must have taken his sweeeeeet time to lose that way in such a short time.

This game makes no sense to me. Neither side had much checkmate pressure (only one check occured), both sides had two rooks, a bishop, and a queen. I've rarely seen such a balanced game. What the heck happened?

Re:I agree, i'd love to see a computer defeat a 9d

[epine]

For this reason, many in the field of artificial intelligence consider Go to be a better measure of a computer's capacity for thought than chess.

Of course, in the history of human civilization in every case we've resolved simpler problems before more complex problems. Human success order is an absolute metric on problem complexity.

Not a good measure of intelligence

[GPS Pilot]

Take an arbitrarily large corpus of writings sampled from the world wide web. This corpus will establish the equivalent of an IQ test. Give the AIs the task of compressing this corpus into the smallest representation. This representation must be a program that, taking no outside inputs, produces the exact sample it compressed. The AIQ of an AI is simply the ratio of the size of the uncompressed writings to the size of the program that, when executed, produces the uncompressed writings. In other words, the AIQ is the compression ratio achieved by the AI on the AIQ test. The reason this works as an AI quality test is that compression requires predictive modeling. If you can predict what someone is going to say, you have modeled their mental processes and by inference have a superset of their mental faculties.

When a very intelligent human studies, understands, and internalizes the writings of others, he or she will indeed be able to model another person's mental process and roughly predict what they might say. But that's not the criteria in this proposed competition. For this competition, the AI is required to reproduce the exact sample that was compressed.

By this standard, I'm afraid that WinZip 9.0 will be judged the world's most intelligent AI.