ARTelevision

People and computers pair off in games of chess occasionally. They've been doing it for decades. Of course, computers are evolving exponentially. Each successive round of Grand Master vs. CyberBrain brings this all into focus.

It's happening again and it's worth our attention, I think.
..........................................................................................

Kasparov 'obliterates' Fritz with strategy

10:52 17 November 03

NewScientist.com news service

World number one Garry Kasparov "obliterated" the chess computer program X3D Fritz on Sunday with a strategy that proved too forward-thinking for the machine to defend against.

"He steamrolled right over Fritz," says John Hernandez, chess consultant for Fritz. "It was a masterpiece." Kasparov achieved a winning position after just 16 moves. But playing cautiously, he did not finish the program off until the 45th move and over four hours.

In the previous match, Kasparov lost by making a blunder a computer would never make. But on Sunday, the position was reversed, with Fritz displaying a weakness that dogs chess programs - being incapable of understanding a long-term strategy.

"Almost from the start Fritz did not understand what was going on and just shuffled his pieces around aimlessly," says Jonathan Schaeffer, in the computer science games group at the University of Alberta, Canada. "Kasparov won effortlessly without giving Fritz an opportunity to do anything."

Commentators made it clear during the match that a human player would have known how to counter Kasparov's play. He used a similar approach to break down program Deep Blue in 1996.

Sunday's result means Kasparov and Fritz are now tied in the four game series in New York, following a draw in game one and the victory for Fritz in game two. "I am in a very good mood," Kasparov said directly after the match.


Inherent weakness


The match went straight to the heart of inherent computer weaknesses and human strengths, says Schaeffer. A human player is capable of implementing a plan that takes over 40 moves to execute, whereas a computer can only take each move as it comes. The program does its best to improve its position, but each time all it has is to refer to is the moves it is capable of probing in advance.

Fritz analyses three million moves per second and uses smart software to probe promising moves much more deeply than less interesting ones. This is in contrast to the "brute force'" approach of programs such as Deep Blue, which attempts to calculate all possible moves, relying on massive computing power to achieve deep searches.

Fritz actually probed deeper this game than ever before, sometimes looking 19 half moves ahead. But nevertheless it still could not rival the long-term strategic skills of the human brain.

Without understanding what Kasparov was doing, Fritz was incapable of a defence. "All the searching in the world couldn't have made a difference," explains Schaeffer.


Closed game


Kasparov also played a very "closed" game by building a wall of pawns early on. This exploits another computer weakness. In a closed game, pieces interact very closely, meaning there are fewer choices per move.

This puts the human player at an advantage, because winning then relies more on having the best strategy. In contrast, computers play better in open games where the sheer number of possibilities to consider often foils humans.

Man and machine will face off one in the final match on Tuesday. Kasparov will be confident after Sunday's victory, but will have the disadvantage of playing black.

......................................................

Besides the thrill of wild abstract mental competition this story evokes, it also allows us to think about the cyberbrains we interact with every day. Can you feel yourself sort of adjusting to the way your computer(s) think?

I can. I can feel it every day. I think more and more like a logically organized device. And I think that's a good thing, in general.

How about you?

__________________
create evolution

Spartak

Go Gary!

Kasparov was my childhood hero, chess was huge where I was born

__________________
I can read your mind... looking at you... I can read your mind...

Bloodslick

I believe that I find myself acting less logically than before. There are small day-to-day things that I am able to approach in a more logical manner, but my actions and emotions are becoming less logical and more bestial. I can leave logic to computers while I regress and enjoy the primitive side of emotion.

glytch

I watched the whole match on ESPN2 (yeah, they showed it, I thought it was strange as well.) The only thing that I kept thinking is that Gary isn't playing against the computer. He's playing against the programmers and damn near every chess strategy ever thought-up. Doesn't seem quite fair to me.

Also, now he will never be able to use the same strategy, or any like it because the programmers will be prepared. In their minds, (and many others) it is only a matter of time before it will become impossible to defeat one of these chess machines. I think, however, that until AI is truely developed the human will always have an edge, even if it is very slight, because of improvisation.

__________________
If you're flammable and have legs, you are never blocking a fire exit. - Mitch Hedberg

ubertuber

What I think is cool is that this computer uses a refinement of software strategies rather than an escalation of computational power. Reminds me of a project where two universities taught those little Sony Aibo dogs to play soccer. One team utilized strategy and teamwork and the other strengthened the speed and reflexes of the individual players. Not surprisingly, the team oriented Aibo dogs won the match. Does anyone know more about what I am talking about?

What would be interesting would be to see Kasparov team up with programmers to write software that imitated the strategic thought process. Then you could run that software on a big computer. I bet that a strategic program would defeat a vastly more powerful machine that used the brute force only method.

__________________
Cogito ergo spud -- I think, therefore I yam

powder

I think the thread title should read "human matches wits with programmer"

Macheath

I'm not really an expert on the logistics of chess, but I have read that it's one of those games that's well suited to the "mind" of a computer. There are other challenging games like the Chinese "Go" that are harder to program to the standards of these Chess programs and wouldn't necessarily be able to challenge humans in the same way.

Maybe someone who knows a lot about programming games can correct what I've written here.

It's a great story but I don't have any arrogant feelings about the human intellect though. To me this is an assesment. This is a way for humans to understand their own "specifications", what is is about the programming of the powerful human computer that gives it certain relative strengths and weaknesses.

Using computers each day makes me want to learn more about them, and new applications for them. I find myself thinking in very systematic terms; persuing more automation and multitasking. This is the case moreso since I'm now interested in learning the best way to set up a decent home network.

I feel less desire for raw computing power and more for systematically "getting the most" out of the setup I've already got. I like thinking in terms of the eventual interconnectedness of all the technology in my house. And, God help me, I'm even starting to cannibalize old computers for parts.

nukeu666

fritz will get my cheers

__________________
Why did the Comp. Engineer get X-mas and Halloween mixed up?
Because Oct(31) == Dec(25)

itchy93

Macheath

Go is next on their agenda.

__________________
That's just my opinion; I could be wrong.

Macheath

I heard something about that and I think it's a really good thing to focus on. I remember reading that the programming a good Go game requires many more years of work without any guarantee that in the the end you'd have a brilliant program. It's been said that researchers perhaps go for the glory of building a great chess program rather than the slow slog of developing a Go program.

Me, I don't even know how to play it but I kind of wish I did.

dnd

Thats not quite true, i've only done a bit of AI but i wouldn't be surprised if complex neural networks were implemented which mimic the neurons in the brain firing. Fritz has no doubtly been taught from scratch by either playing humans or other computers although i wouldn't be surprised if these had been pre-programmed with strategies.
The idea behind deep blue and Fritz is not about the programmers winning but displaying the improvements in computer processing and AI techniques....

(i may be totally wrong!)

__________________
"Never underestimate a dumb question"-- Brandon Boyd

Gortexfogg

I find this whole thing very interesting. Kasparov not only has to think like he's playing a human, but also like he's playing a computer. It'd really be interesting to hear more of what he says about battles of wits with computers.

AI will probably know more and be able to recall information much faster than us, but I doubt it will be able to think the way we think, unless it's possible to replicate emotion and such...

glytch

On the 6th or so move, Fritz made a very unorthodox move that completely thre Kasparov for a loop for a good 6 or so minutes. This, I think, is the number one advantage of the AI opponent. You can't throw it off guard. In chess championship matches, a large part of the victory is psychological, which is the biggest problem when playing against a computer.

__________________
If you're flammable and have legs, you are never blocking a fire exit. - Mitch Hedberg

Fire

part of me is intrigued by this contest, as it represtent the evoloution of computer AI, despite this i realy want the human mind to triumph- the reverse is spooky

__________________
Thought the harder, Heart the bolder,
Mood the more as our might lessens

j03th3j3w

its the classic man vs. machine... you cant help but bring up the matrix... i had to... AI will eventually get smarter! and it will become more beneficial to our society... but.. not matrix style.. to where we all become ectasy rolled out ravers in some cave in the earth! not that far...

TIO

To all those who say Kasparov is playing the programmers; that's not quite right. The cool things in computing these days are done using adaptive systems, which are systems that develop themselves over time. The computer would have learned how to play the game; it's not going through set algorithms the programmers have created (Glytch suggested that Kasparov was up against 'damn near every strategy ever though up', which is far from the truth. He's only up against the ones Fritz has learned and developed himself). Yeah, the system will only be as good as the programmers who wrote it, but the programmers don't have to be good at chess for the program to be great. They just have to be good at writing adaptive systems. I'd wager that Fritz could absolutely thump every one of the programmers who wrote it, every time.

dnd, I don't know much about the architecture of Fritz, but it's unlikely that it is a neural network. At the moment, they're not very good at chess (next year, I'll possibly be working on a neural net that plays checkers, and even that is pretty challenging). More likely, it'd be some kind of genetic algorithm.
(They've also done some nifty things recently with game-playing cellular automata, but they're still a long way off world-class)

As for Go, that's the holy grail of computer gaming agents. There are way too many possibilities in the game for the computer to search even a fraction of them. The game seems to be played more by intuition than strategy; even great Go players can't really explain how they play well. So what you're trying to teach the computer isn't intelligence, which they're fairly good at, as much at it is intuition and judgement, which they have quite a hard time with.

__________________
Strewth