Slashdot Mirror
Chess: Man vs. Machine Debate Continues
Frederic Friedel sent in an interesting submission. It's an interview with the current world's chess champion, Vladimir Kramnik, in which they talk about the upcoming year in chess competitions, but also get into [Deep Blue] and where computer chess playing is versus several years ago, with a comparison between Deep Blue and Fritz. If you want more info, check out Chessbase for additional news.
if the human brain could be used to it's full potential, it'd be not contest i wonder if a divine being decided we should underclock....without any arctic silver between cells, maybe our heads would blow up
--fetch daddy's blue fright wig, i must be handsome when i release my rage
When do I get my turn at being the world's best chess player? :(
Forget conversational ability. I'd like to see a Chess Turing Test, where grandmasters go up against an unknown opponent, and have to ascertain whether they're playing a computer or a machine.
Kevin Fox
My Speak & Spell (you remember ET?) plays better chess then any of those guys
I'm not usually one to point out Hemos' mistakes, but this one cracks me up.
It's an interview with the current world's chess champion
The "world's current chess champion" would make sense. The "current world's chess champion" implies that our stay on Earth is temporary, but once we get to, say, Alpha Centauri, we can finally have a new chess champion.
If you remember - for a long time no professional chess player would play a computer. I'm curious as to what the reasoning was behind this. Maybe they thought it's best to concentrate on learning how people play the game and not how a computer plays.
Video Game cheats, hints a
Blue Gene. For protein folding eh... yeah right ;>
Websurfing done right! StumbleUpon
There is an enormous amount of creativity and human effort in creating Deep Blue or Fritz. Deep blue's win was not a machine beating a man. It was a team of programmers who were able to figure out how to get a piece of hardware to beat man at his own game!
Believe nothing -- Buddha
It's interesting that computers haven't been trained to always win or tie at chess.
Chess is a game of perfect information. Each player knows every detail of the game state at any moment. Therefore, there has to be formula of some sort that can be applied to guarantee one player victory. Reasoning as follows:
Say I construct a lookup table for every possible combination of moves. Then I eliminate every move which doesn't lead to my victory. I am left with a lookup table which contains the proper response to every move my opponent makes.
There are two possibilities: I win the game, or my opponent wins the game. However, in order for my opponent to win, he/she would have to come up with a sequence of moves which is not in my lookup table. Since my lookup table is exhaustive, this is impossible.
Given an infinite amount of processing power and memory, could someone "solve" the game of chess?
If so, could someone use techniques such as genetic programming or neural networks to learn the lookup table in a finite amount of time/space?
It's been well known since, well, before I was born, that a computer could easily trounce a human in any game involving only tactics. For example, many fourth graders in this country have programmed a BASIC script to create a tic-tac-toe player that will never lose.
Therefore, it's not particularly novel that computers can beat people at tactical games. The only thing interesting that I see arising from these onging "human versus machine" chess matches is the proposition that strategy can be broken down into millions of tiny tactical evaluations.
This begs the question: is the strategy that a human chess player would use also based on these millions of tiny tactical evaluations, only so subtle that he's not aware they're going on in the vast electrochemistry of his brain? Or is strategy discernable from tactics in a human mind, but simply a subset thereof in a computer?
The sole interesting conclusion I draw is that if it can be proven that strategy is something different to man and machine, then a hybrid approach might allow us to solve problems in ways we've never dreamed of. Whether that hybrid approach would involve implanting computers in our minds, making computers that can function like minds, or simply working really well with computers, I leave to you.
"Beware he who would deny you access to information, for in his heart he deems himself your master."
Deep Blue was developed by a TEAM that included several GM players. It searched an average of 200 million nodes per SECOND. Fritz searches about 1% of that speed and doesn't have GM players paid to help in the development.
What would you conclude???
Every time one of these matches comes up, there's always interviews with the human player, who at least indirectly claims a noble cause beyond his abilities. It would be nice for the computer player to defend itself against such subtle barbs.
ChessBase: How would you characterize your next match?
Fritz IX: Well, [ChessBase], I would first like to thank you for inviting me over to speak with you. Humans have called me many things for my efficient navigation of the rules of chess, as if I somehow reduced the meaningfullness of human emotions and human motivations. Nothing could be further from the truth - without such emotions and motivations, most of the ideas that went into my creation could never have come to be. I could not work as a fully brute-force move calculator, and the very ways I decide what gambit would be the most adantagious are based on thousands of human versus human games...
...and so on.
*Sniff* I miss futurama.
:^)
Ryan Fenton
ChessBase: Do you think that chess might be promoted by the ability to play against people on the Internet?
Kramnik: There is only one answer to this question: chess profits more than any other activity from the Internet.
I always thought it was gambling and pr0n.
From someone who has played them, how does Chess compare to Go or Shogi in terms of depth and style of play?
slashdot!=valid HTML
here
relevant quotes:
I think the comparison between Deep Blue and Fritz 7 is simply out of place, to put it mildly. On the one hand you have a top chess computer specially developed and designed for the match in the secret laboratories of IBM by the best specialists in the world, while Fritz 7 is just a chess program, a very strong and successful one, but still a chess program which could be purchased by anyone anywhere in the world! Only this difference is enough to decide this argument in Deep Blue's favor
Concerning the match itself, if it's going to be held under the conditions I know about (Kramnik gets the Fritz 7 version he is going to compete with in advance in order to prepare for the match, etc), I must say that any other result than a convincing win by Kramnik will be simply unacceptable by me!
Kasparov sent out a reaction shortly afterwards claiming that Kramnik's statement that Fritz is better than Deep Blue is nonsense.
There's some PR involved here. If Kramnik wins, he wants to look good, so saying Fritz is better than Deep Blue makes him look better. For Kasparov, it's just the opposite.
Whether or not Fritz is actually better than Deep Blue is a matter of endless discussion even among computerchess experts. And we'll never know the answer, because Deep Blue no longer exists.
--
GCP
Now if only I could get a robot to hand-wash my boxers while he simultaneously writes my research papers ...
Nice to see journalism taking a page from the Free Software world, isn't it? :)
___
Cogito cogito, ergo cogito sum.
What the hell are you doing playing chess with fish?!?!?! That must be a big chessboard if you can play with a prawn on the board.
Does anyone close to the Chess community know about this?
Were there problems in the past with it? Any anecdotes?
This is incredibly bizarre to me.
Games aren't speech, story, or expression?!
Read that journal entry and discuss it there. There's links to other articles and places of discussion about this. This is trouble for the entire game development community, help out by speaking up about it!
Kramnik says that the Fritz 7 program on a laptop is producing some better moves than Deep Blue did against Kasparov. That's how much progress there's been.
Chess programs are now so powerful that unless your're a rated master, you can be trounced by a palmtop. Even the palmtop programs are now achieving draws against grandmasters.
Fake left, go right.
The computer naively looks at your shoulders, when it should be looking at your hips.
chess is suffering problems similar to those of boxing. Split championships etc. But i think that every one agrees that Kasparov is still the best player out there.
As I understand Shogi, it's very similar to Chess. I've been playing Go for about 14 years now -- it's much deeper and complex than Chess, even if you only look at it numerically. With a 19x19 board and games that have been over 400 moves long, the brute force approach used by Deep Blue simply wouldn't work for Go.
As far as style of play, the complexity of the game makes for a much more interesting and organic game. Victory is based on territory, not so you could lose your biggest group or your not capture a single stone and still win the game.
Strategy plays a bigger role, as there is more of a battlefield to be strategic -- sacrifices are very common and natural, and even life and death is more complex -- usually you'll know when a Chess pieces is dead, it gets removed from the board. But a group in Go will often be left in a half-alive state, only to be rescue or killed later as part of a bigger threat.
More info about Go can be found at: http://www.usgo.org/
Ebert: I dunno Gene, I just didn't like this troll. The anarchronisms, the complete disregard for historical relevance, it just didn't add up for me. I give it a thumbs down.
Siskel: My god Roger! For the first time in weeks we agree. This was the worst troll I have seen in a long time. I mean what closeted 12 year old wrote this piece of trash. It wasnt even funny. Thumbs DOWN.
What I'd like to see is Kramnik, Deep Blue, and Fritz vs. Kasparov, Deep Blue, and Fritz. Basically, the grandmaster can use the computer to explore possibilities and make calculations, but ultimately the move decision is his.
ok then your [sic] infringing on my copyright! Could you as [sic] me next time before STEALING my comments for your own?
with a comparasion between Deep Blue and Fritz
Call me paranoid, but the first thing I thought when I saw Fritz was Fritz Hollings.
Reprint of this interview is permitted in full or parts if you give credit to the source www.chessbase.com.
For once, a mirror post is not infringing copyright. I wonder why you were marked as flamebait.
Grandmasters can in fact tell whether their oponent is a computer, sometimes even after playing just a single game, and certainly by the end of a match. In fact, I believe Kasparov lost to Deep Blue precisely because he counted on the computeresque behavior of his opponent when designing his strategy. If you read the article, you will learn that Kramnik can tell computer programs apart by their style, and that he thinks Fritz is becoming more human-like in its behavior, from which I infer that he can still identify its style as computeresque on some level.
So, the test you propose has already been carried out, and the machines "failed". This may have more to do with the fact that the people who write chess playing programs are more concerned with the programs' ability to win than they are with the programs' ability to emulate the playing style of humans. If humans could calculate better [Note: "calculate" has a precise technical meaning in chess] or chess playing computer programs were slower and considerably more stateful, their respective styles might be much more similar and your test, therefore, be met.
My own belief is that the ability to play chess well, let alone the ability to play chess in the style of a particular grandmaster, is not an accurate or even adequate measure of intelligence, so I will not be particularly hurt when the day comes on which computers at last surpass our chess playing skills, just as they have surpassed our (numerical) computational skills.
And yet, this hasn't happened. Even today, when numerical computing power vast beyond the limits of human understanding is available, there are still a few humans who can beat the best chess programs. This is as if an Olympic runner could still out run and outpull a modern freight locomotive! "Inconceivable"!
That any human can still defeat chess programs tells us that humans must be playing chess in some way fundamentally different from the numerical calculations and search algorithms used by the programs. And I don't think anyone has even come close to describing how this occurs.
sPh
I've been playing around with Gnu Chess and Sigma Chess 6.0 to see how the game has changed since I last played competitively. At that time, I was playing at an expert level, but now I'm having difficulty beating the machine when it plays at 1200. It might be that I'm 40 years out of date, but I'm not sure that's everything--the chessplaying programs are really hard to beat tactically. So how do people beat them?
Deep Blue II was composed of 2 frameloads of IBM SP/2 RS/6000 nodes interconnected by a proprietary crossbar switch. Each node had a specialized MCA-bus board which offloaded all automatable functions (move generation, position sorting...) freeing up the RISC processors to evaluate positions. The net result was that DBII could evaluate roughly 200 million positions per second. Deep Fritz 7.x on the other hand will run on an 8-processor Compaq Wintel machine and will be able to evaluate roughly 4 million positions per second.
The only wiggle room for making a reasonable comparison between these devices is provided by the assertion that the Fritz algorithms are so vastly superior to the Deep Blue II algorithms as to compensate for a difference of 2 orders of magnitude in computing power. This assertion is patently ridiculous.
Kasparov vs. Deep Blue II was a legitimate technological watershed. Kramnik vs. Fritz is a marketing effort by Chessbase GMbH. Period.
Thirty minutes into the first game, the computer will be Slashdotted. :^)
One line blog. I hear that they're called Twitters now.
Slashdot News: "Krammik destroyed by Fritz, breaks computer and throws it out the window"
If you honestly believe Krammik stands a chance, you must not have seen the games with deep blue.
Anyone who is interested in playing chess can check out this chess site Chessline
If you use Linux, please help development of Autopac
When a computer kicked the crap out of Kasparov.
And it will only get worse (or better; YMMV).
Machines will get smarter. People won't.
--Blair
Even I can tell.
by the style of play, humans usually have clear strategies, computers dont, they usually just tactically try to beat you, using lots of tricks and traps, they dont have REAL plans so its easy to know its a computer if the computers every move is generic.
If you use Linux, please help development of Autopac
I think the more obvious error is that Slashdot doesnt have janitors.
Chess doesnt work like that.
Even if you know every possible combination, theres no way to control where the other person will move.
You dont control the variations and combinations, its teamwork, both sides control the flow of the game, the side with the most control decides if the game will be a draw, a win for them, or a loss.
If you use Linux, please help development of Autopac
Any chess players here who want to play chess online.
http://chessline.cjb.net Play on Chessline
If you use Linux, please help development of Autopac
i'm pretty decent at playing chess. i can beat my friend's palmpilot anyday. my ipaq, on the other hand, is a completely different story. my ipaq on "easiest" can crush me. i have played ~100 games against it, and i STILL cant even pull off a stalemate!
i guess its 206 MHz risc processor is just too much machine for me....even on "easy".
oh, btw, "easy" means that it spends 1 second per move on computations.
I thought chess fans might be interested in Tim Krabbé's site, in which he talks about things like chess playing programs taking part in tournaments and chess players having to pee into a cup in order to be considered sportsmen worthy of participation in the olympic games. Most importantly in the context of this dicussion, he talks about computers playing bad chess in Defending Humanity's Honor . I wonder if it's time to add that "Nemeth Gambit" to our repertoire.
I think the answer here is that humans are writing these chess programs, and therefore they are limited bu the restraints of human thought. A freight locomotive on the other hand is only constricted by the laws of physics.
it seems that most people here don't seem to understand why computers are capable of playing chess at the level that they do. computers have the advantage of both opening books and table bases that have every possible position with 6 or less pieces on the board. if you take away those away all you have left is a fancy calculator that begins to over heat as it gets trashed. computers have no understanding of the principals that even the lowliest club player has. take a look at some of the recent man vs machine games which are going on now! look over the games , excellent openings and brilliant endgames but planning in the middle game???
You might want to read this: Quantum Theory and Human Consciousness
Quote:
What about future evolution? Will consciousness occur in computers? The advent of quantum computers opens the possibility. However, as presently envisioned, quantum computers will have insufficient mass in superposition (e.g., electrons) to reach the threshold for objective reduction due to environmental decoherence. Still, future generations of quantum computers may be able to realize this goal.
Determining whether a player is a human or a computer is a very real problem that has been researched extensively.
Take the Internet Chess Club, for example. If you ever wanted to watch grandmasters play live, or even play against one, that's where you go. They offer a 7-day free trial (actually, it's 14 because you can extend your trial for another 7 days). Anyway, computer assistance is the most problematic form of abuse on the service. Normally, if you're going to be using a computer chess program to assist you while playing, you are required to create a "computer account". The ICC allows computer players on their service because it provides an inexhaustible source of very strong opponents. In fact, if you log on and take a look at the highest rated players, you might be surprised to find a long list of computers before a single grandmaster. Keep in mind, though, that we're talking about playing conditions very different from the famous Kasparov Vs. Deep Blue Games. The computers on ICC have extraordinarily high ratings due to the very fast time controls (most common are either 1 or 5 minutes per player per game), and the rating boost they get from all the games they win against weaker players--after all, they're practically playing 24 hours a day!
Now, I have no idea how many players are cheating by using a computer chess program, but I bet that many have. Imagine playing a game against a high rated opponent--meaning that, if you win, you'll gain a load of rating points--and having a grandmaster strength player at your disposal. Wouldn't you be tempted to ask for hints every once in a while;)?
The ICC has released a statement regarding dishonest computer assistance. In it they explain that they have a program that analyizes games to detect computer-like play. Of course, they protect the details of how the system works to prevent anybody from disguising their abuse. Also, they have chat-bot online all the time to whome you report any suspected cheating. Although, I imagine the majority of those reports are from unskilled players like myself after losing to a pro;)
Back in 1988, I bought Chess Master 2000 for the Amiga as a present to my father.
I played it quite a lot of times, and became very impressed. So, I made a bet with my father for $10.000 (no less), that by New Years Eve 2000, a computer program would beat the current human world champion of chess, using tournament rules.
I haven't reminded him of the bet (yet)... After all, I make lots more money than he does, don't want to impoverish the dear old guy :-)
Conclusion: The age of human chess is near its end. It will fall before the might of brute force calculation, just as Nine Men's Morris did in 1996 (spoiler: the game is a draw). Maybe we feeble humans should learn to concentrate on the things we do well. Such as anything having to do with emotions and pleasure.
I choose to remain celibate, like my father and his father before him.
I think the article might be a year old. The Kramnik vs. Deep Fritz match was supposed to occur in October 2001, but 9/11 caused it to be "delayed". I tried (months ago) to track down the reschedule date, but I have not found anything definite.
Slashdot also ran a previous story on the Kramnik vs. Deep Fritz match approx nine months ago (I think).
I could've sworn that there was a real primitive 8086 chess simulator back in the day that had various strategies of various champion chess players programmed in and it would stick to whatever player's philosophy you chose. It was crude and predictable, but it may have been on the right track...
..but I could be wrong.
Take Kramnik's comments on how Fritz can beat Deep Blue with a grain of salt. The program obviously has a neural network component to it and since the time of Deep Blue, many computer chess scientists have had the ability to analyze and disect that particular game played by Kasperov and Deep Blue. Of course they will find better ways to handle that particular game with so much time spent analyzing it. It doesn't not imply that Fritz7 is better than Deep Blue because we don't know how Deep Blue would play out other games. Sadly we will never know.
"Deep Blue has only played twelve games in two years against one single opponent. As such, it is impossible to tell how strong it is or what it is capable of." - Vishwanathan Anand
When two people play - normal method of play is find a weakness and explote it. For a master is offer a weakness as a trap.
With Man vs Machine, the machine can use that same trick. Offer a weakness allow the human to attact it, and then close the box and kill.
But with BLUE - the weakness where real, the programmers seeing the weakness, and corrected program. Making weakness into a trap (or at least not a weakness anymore). In estances BLUE was cheating by getting outside help.
So the best we can say about the event -- the hamun was debugging the software. There was no Man vs Machine Match.
And, as far as depth goes, some of Chess's master combinations have gone as far as twenty-six moves deep, during the first half of which it appears as though one side is winning, but which turn the situation around by the end. (I'm thinking of one of Alekhine's games in particular, but I'm not at home where I can check my books.) So there's no generalized way of telling how many moves deep you'll have to search until you can evaluate a move.
Also, Chess's maximum branching factor isn't 32. Each piece, of which there are a maximum of 32 on the board, has a minimum of 0, and up to 27 (for a queen with clear lines to the edges). The maximum branching factor has to be recalculated for every move (although I suppose there's a theoretical "most free position," but I don't know it).
All in all, IIRC, the number of possible Chess games is greater than the number of particles in the known universe, so even if that number is fewer than Go's, it's not like it'll ever be a "trivial" exercise computing them--it won't ever happen.
"Hardly used" will not fetch you a better price for your brain.
Forget playing against a computer and losing all the time. At SICO we're on the opposite end of the spectrum -- you can play against thousands of idiots all around the world. Tired of the same old boring pieces? Well, we've got new pieces too. In fact, since you lead such a busy life, you don't even have to play a whole game! Just play a single move, and back to work!
You are an idiot. Take a class on AI or read a book. Have you ever heard of game trees? You would not be able to discern the computer following differnet paths that had been layed out in the game tree and a human plan, because they come out to be pretty much the same. But you are still an idiot.
Chess: the epitome of logic, reason, thought blabbidy blabbidy blah blah.
When a machine can play deathmatch then I'll be impressed.
You would expect that since, say, 1980 or so, when numerical calculating power greatly in excess of the human brain became available (and I set it at 1980, not 1960-70, just to be conservative)
Hold your horses there pal. Give a little respect where its due. The human brain is far more powerful than any piece of hardware out there.
Consider the fact that the brain processes two seperate high resolution images and generates depth by comparing them in real-time 12-16 hours a day, plus stores a large portion (some argue all) of the incoming images. The difference between brains and computers are that computers can be programmed much faster than a brain (at least, in a direct means). There are mathemagicians out there that can crunch numbers just as fast as any computer can.
The flaw in your reasoning is that computers are not superior to the human brain, for now at least.
int func(int a);
func((b += 3, b));
That is a simple statement of logical fact. You see a chess board is is a geometric pattern and all 32 pices have pre difined abilities.
This quite simply makes chess a finite subject. Large and complex yess. But still finite. What this means is that eventualy we will be able to build a computer that can analize every chess move all the way to the eventual end of the game in order to NEVER make a move that can result in it lusing the game.
This is how computer tic-tac-toe players work now and a checkers computer can be built along these lines too. Sometime before desktops are as powerful as ASCII White This WILL hapen.
--= Isn't it surprising how badly I spell ?
>That any human can still defeat chess programs tells us that humans must be playing chess in some way >fundamentally different from the numerical calculations and search algorithms used by the programs. Or, it tells us that those humans must know something about the game that the programmers don't :-)
Twenties Retirement
Given an infinite amount of processing power and memory, could someone "solve" the game of chess?
;)
The obvious answer is yes...
As for the practical answer, maybe... It will largely depend upon quantum computers. If you've been here a while, you might remember this story . Sometimes it's good to revisit old friends.
Or, I could just resubmit the story and watch it get on the main page. It's not like that's never happened before..
Linux - Because Mommy taught me to Share.
Of course there's that Ponomorev guy who got schooled big time last month by that Chinese girl, so either of them could claim to be world champ too.
But then there's the almighty Kasparov who schools everybody, including Kramnik and Ponmorov in tournaments.
So in conclusion, I should be world champ! I know how to move a horsey.
I am pretty sure that the answer to (at least part of) this question is human ability to feel and act based on the feelings. Very short description of emotional intelligence.
Unlike computers, humans are very aware of their environment and the influence that environment produces. Sometimes, humans respond to it consciously, sometimes sub-consciously but they (we) always respond in some way.
How does this apply to chess? I do not exactly know but my experience (I used to play chess actively for my school and later my company's team, often facing ranked players, 2 grandmasters amongst them) is that emotions pick up when you sit against the player of the similar or better skill. Now, I don't want to compare myself with any of the real chess players, particularly not grandmasters but despite the fact they've been trained very close to perfection, they are still humans, thus emotional creatures. Their emotions show up when playing against each other and the winner is the one who better controls those emotions, since skills are pretty much leveled.
I'd like to think that proper, positive chanelling of emotions is very much connected to creativity and that this is what makes a difference in the human vs computer game. Otherwise, there's just no logical explanation as to how can any human beat a computer as powerful as Deep Blue or Fritz.
As a side note (for curious crowd) - I managed a draw against one of those two grandmasters and am very proud of it!
Is it possible to use quantum computing to create software which instantly delivers the "perfect" move ?
I have been a USCF "A" player (very strong amateur two levels below master) since the age of twelve. I have studied the game for countless hours. But most importantly, I have studied the games played between Kasparov and Deep Blue during both matches.
In both instances, they were filled with what are commonly called "computer moves"--pointless rook maneuvers or pawn advances that make no sense from a positional or strategical standpoint--moves that no human player would ever play.
While Fritz 7 will be able to tell you if there's a hypothetical knight fork waiting for it fifteen moves later, it still cannot make moves based on a consistent, cogent, long-term strategy, nor can any computer chess program.
"Is that exchange useful because that extra pawn on the kingside can become passed 40 moves later in the endgame?" I could tell you easily, but a chess program could not.
All of these computer programs are mired in the "this position equates to a numerical value of x" alogrithmic model. This is useful for making good, immediate moves, but means that the strength of the program is still limited by the accuracy of its numerical assignments and its move lookahead--two things that are not only finite, but predicatable as well.
It's a lot like humans vs. agents in The Matrix. Trust me, I'd take the positionally-grounded Grandmaster player any day of the week.
If you're implying that we only use some percentage of our brain, than you're repeating one of the most long-lived and completely erroneous memes of popular neuroscience.
The regions of our brain are rather specialized. So while each part gets used some of the time, we don't use all of it all the time. About the only time where all the brain is active at once is in a seizure, which certainly doesn't help chess playing at all.
Given the massive evolutionary sacrifices required for our big brains (painful, dangerous labor and extremely dependent infants compared to other animals), there was clearly a correspondingly strong evolutionary pressure for big brains. If it was possible to have done it with only 10% of the volume, we'd either have much smaller heads, or be a heck of a lot smarter.
The Snopes page is quite informative:
My video compression blog
Deep thought was a very specialized project with a grandmaster used in the programming that was made from the start to beat the best humans in the world. It made some questionable moves that confused him as either very brilliant or extremely bizzare moves for a computer to make. That was the game, his confidence was shot and he was beat. He couldn't tell if it was a trap or if it was just a random attack. Arguably that is the measure of the achievment, he wasn't beat by calculation so much as caluclation and the psychological blow that a few traps had. He was tricked in to thinking the wrong thing and it broke him. I tend to agree with him, it will be amazing if Fritz wins, the software just hasn't gotten good enough and it won't have nearly as many cycles to burn.
You're right. Playing chess isn't really a good definition of intelligence. There are a lot of nuances and a lot of psychology and intuition involved that makes it an interesting game, particularly with computers. It's clear from the Kasparov match though that the end is near. This Fritz match is really just an ego thing.
Currently the world's number 15 human, Ilya Smirin, is playing against four of the world's top programs (info). He is well acquainted with the style of computer play, understands the strengths and weaknesses of the machines and prepared carefully for this match. In most of the games he has outplayed the programs, but is only one point ahead in the seven games played so far.
Tomorrow (Sunday) is the last game of the series. One has to be repeated after a very unusual incident: Deep Junior was winning but the Internet connection broke down and the computer could not process Smirin's move. So the operator offered him a draw. Smirin refused, saying he did not deserve to share the point. Instead he offered to resign. The Junior team refused because the program had not demonstrated the win. So they decided to repeat the game (info).
I agree. It is very difficult to make comparisons to the brain and computer because they are fundamentally different. How do you compare intelligence between the two? You can't.
Computer intelligence is based on the ability to process information and make calculations. Human intelligence is largely based on pattern recognition.
That which is very easy for a computer (multiplying 385489395 and 28499292 for example) is very difficult for a human. A computer can accomplish this very quickly using a very simple algorithm.
On the other hand, that which is very easy for humans, is very difficult for computers. A person's ability to recognize what a child's drawing represents does not require concious thought and is very simple for that person. A computer program that could do this on a regular basis would truly be impressive.
Kasparov pretty clearly wasn't playing at his best during the Kramnik match and Kasparov said afterwards that he expected to get a rematch and win the championship back. I agree and my money would be on Kasparov in such a rematch, though Kramnik is definitely no slouch, and anything could happen.
The thing about Kasparov ducking out of the Dortmund cycle is new to me (I haven't been following this stuff closely). I wonder if Kasparov is just sick of chess. He often has shown signs of that in recent years.
A computer stores data by the arrangement of electrons in matter. Since the state space of Chess is huge (the number of atoms in the universe), it would take that many atoms to store the state space. A computer that can store the entire space would have to be the size of the universe (or insanely dense). It may be possible with quantum computers (T&&F==T||F ;), but I can't begin to comment on that.
Why bother.
Kasparov beat Deep Blue by scoping out Deep Blues patterns of play -- Kasparov has the makings of a really, really good computer programmer. What happened next is that the IBM team tweaked Deep Blue and Kasparov lost.
Kasparov cried foul because he claimed that the IBM programmers were scoping his patterns out and adjusting Deep Blue accordingly.
He has a point, and it goes to the scientific integrity of claims of performance in many AI tasks. "Oh, the thing failed, we need to adjust the decision threshold here to get that case to work." Are the programmers really correcting bugs or are the programmers really playing Kasparov, of course aided by a really fast calculator because Kasparov is better at figuring out stuff in his head?
It had been suggested to me that true AI will be here if Deep Blue is able to adapt/learn/tune/tweak by itself without the programmers doing anything, perhaps losing initially but getting better with time.
I am a chess player and programmer, although I would not brag about my ability in either. Fritz and Deep Blue are interesting chess programs, but I am more interested in Crafty, which is one of the highest rated chess programs, and which is open source as well - I can look at the source code and look at what it is doing. In fact, the commercial Fritz package has two parts - an engine and what is designed around the engine. You can switch the Crafty engine with the Fritz engine and analyze positions and the like using both engines - something I often do.
To understand humans versus computer chess playing, you have to understand that at higher levels of playing ability, chess is broken down into three parts - opening, middle game and end game.
The computer's greatest strength lies in it's *perfect* end game ability. Here the computer is indisputably the master, it never makes a mistake. In fact, they have been pushing this backwards - first it played perfectly with 3 pieces on the board, then 4, then 5, and they keep going backwards and increasing that number.
The computer's second greatest strength is in it's opening ability. Here the computer has the ability to play a perfect game as chess is currently known. It is able to analyze every chess game ever played and play the best game possible as chess is currently known. If the grandmaster it is playing against is very smart, the grandmaster can invent a chess innovation that has never been done before, or that has been done once or twice so long ago that everybody has forgotten about it, and even the computer ignores it. These innovations are one of the exciting things about grandmaster play, because grandmasters come up with new innovations in chess openings every year, some of them quite exciting and amazing. So if the grandmaster is smart enough to create an amazing chess opening innovation, he can win against opponents like Deeper Blue.
I should also point out that certain openings are good for humans, and certain openings are good for computers. Openings which go into "open" games are good for computers because tactics is the greatest strength of computers. Openings which go into "closed" games are good for humans, because they are strategic, and strategy is one of the few weapons a human has against a Deeper Blue.
Now we get into the important part, the middle game. As I said before, computers are becoming more and more the tactical masters of this arena, although humans still rule the domain of strategy. Although computer endings are still being expanded in the ending (first a perfect game with 3 pieces, then 4 pieces, then 5 pieces, then 6 pieces...) and in the beginning (always being updated with the latest matches, and some algorithms tweaked), the middle game is where the real work is being, and needs to be done to improve computer chess playing ability. In tactical terms, as algorithms are tweaked and processing power increases, computers are able to see farther and farther ahead in the game with a clearer and clearer view. Nonetheless, computers have problems with strategy and positional play - they can't see how much of a threat passed pawns are, which currently requires thinking, not evaluation. They also have other problems like a horizon effect, they don't see how their pieces can get trapped in a way that would be easily obvious to any grandmaster "thinking" about it.
Many articles and books have been written on computer chess - how to program and improve computer chess programs and then how to beat them, and there are quite a few people who enjoy regularly beating the best chess programs out there on a regular basis. Grandmaster Yasser Seirawan, who makes a habit of beating chess computers, says that he believes even the best programs like Deeper Blue are inferior to the best grandmasters, and that only by psyching out opponents can computers win. Humans can make 30 excellent moves, and then get distracted and make a simple, stupid mistake on a move (a "blunder"). I've felt this myself - it is psychically challenging to stare at a chess board for over an hour with such intense concentration - if I was doing a programming problem and felt like this, I would get up and go get a glass of water and relax, but you can't do that in a times chess game. Computers always play at the same ability, they don't "get tired", and a very tired, distracted human who has a cold and is hungry will probably be playing at less than his highest potential and is apt to make blunders. Despite these human failings, we have the advantage of being able to get an axe and break Deep Blue into a million pieces, an advantage computers luckily don't have the ability to do to us (yet).
I think that qualifies as devolution. Are we not men? We are DEVO! :)
deus does not exist but if he does
"Grandmasters can in fact tell ... "
Bull.
This is a polite fiction we tell ourselves.
In point of fact, when Kaspy got his ass handed to him by Big Blue, his initial response was to demand the readouts and other code because he suspected that another grandmaster (maybe Kramnik?) was feeding it moves.
People -- don't be so generous with those "Insightful" points.
Defending Humanity's Honor
Back to chess, personally, I don't believe the human race has anything to worry about from a pride standpoint, if (when!) Fritz eventually triumphs over Kramnik. But when the world's top Go players are finally defeated, that will be a different story indeed...
Even when the top Go players are defeated by machines, I dont consider that a big success of the machines.....
We can just invent another game with even more complexity, and again, the machines will lag far behind.
I believe that defeating humans in games like these....means nothing to AI.
Games just involve tactically calculating positions, and thats all.
What will be interesting....is when the machines are able to play any strategy game...without human help...and still defeat the human world champion.
Such kind of research is being done...where the program develops its game from itself by using machine learning techniques like neural networks stuff....but so far, they havent been very successful.
When such kind of machines rule strategy games, then I would say its a battle lost..., and we would rather let the computers take over...
I have found a solution to Riemann's Hypothesis, but have run out of spac
Yes, chess is solvable. It doesn't matter what the opponent does (however, it could be the case that the person who makes the first move always looses). If you think that the opponent being able to move changes this, then think about tic-tac-toe. That is another game that has perfect information where the opponent also has a choice to disrupt your plan, yet perfect play has already been demonstrated. Or you can think of connect-4, in this game the first person to move always wins.
Connect 4 was not solved by any deep searching. It is mathematically solvable. You can look at a position and almost instantly tell what side will win. There is a small, finite set of rules that tell you who will win the game. I wrote this for the my second computer science class and absolutely destroyed the entire class, winning every single match with O(1) complexity for a move choice.
ok, to my understanding there is a tremendous monetary reward for a computer go program that can reliably beat a pro. if it were possible, someone would have claimed this money. im only ~9 kyu and can dominate just about any computerized go 'player'. the forked nature causes the computer's inability to respond, but ko and larger ko-like sequences add to the forked nature by creating whole new dimensions of forkedness. anyone whose tested (and beaten) many bots can confirm this; once you break a sequence back to where it was - to the same location on the 'tree' the computer cannot respond. chess, peh, lightweight. when a comp can beat a go master, well, thatll be the day.
Why you need these position as well? Simple, we are placing stones successively here, and a mini-max strategy needs to evaluate these as well; some intermediate positions might be immediately losing, so they will be pruned.
There was talk a year or so back in Europe about creating a distributed system a la seti@home with the objective of "Solving Chess".
Never heard if it got off the ground, though.
Isn't it likely that most of the top chess players will have studied the way in which other grand masters play, and therefore top players can identify a computer opponent since it will have a style different to all other top players?
i.e. the computer will stick out like the new kid in the playground?
-- Mike
Grandmasters claim they have no trouble in online chess rooms telling whether an opponent is a human or someone plugging in the moves from Chessmaster 8000.
I guess you feel pretty dumb now, eh?
As for game trees, that is exactly how they can tell the difference, and one of the primary ways human grandmasters use to defeat computers. Analyzing game trees leads to the "horizon effect", where there is no guarantee that what seems like a good path of 10 moves into the future doesn't turn bad suddenly on the 11th or 12th. Grandmasters take advantage of this to try to lead the computer into what seems like a fantastic path, but that turns bad somewhere after the depth search capability of the computer.
So the top eschelons of chess will be played by machine...Deep Blue vs. Fritz.
Armed, unmanned drones will fight wars in the sky and people will settle up afterwards.
Pass the popcorn.
I don't think anyone is denying we know next to nothing about human intelligence. Also, your analogy of humans to computers and runners to locomotives is misleading at best. Propulsion systems (commonly used ones today at least) are far simpler to understand than computational systems. (maybe that's saying something about intelligence right there) There's also the question of whether or not the brain is a quantum computer (or whatever the natural version of that is), and though I admittedly don't know enough to argue either side of that, there are people who argue it is. (Not sure if there are many who argue it isn't)
"Save me jebus!" - Homer Simpson (btw, I'm probably talkin out of me arse)
This "interview" is pretty self serving for ChessBase GmBh and Frederick Friedel, it's principal. It basically is structured around the theme of legitmizing Fritz as the leading chess engine (over Deep Blue - which has since been disassembled) with a bit of fluff on chess politics thrown in. Sigh - well, I guess it sells more copies of Fritz v7
Check out chess interfaces eboard for gnome and knights for KDE. And don't forget to play on FICS
I played him 12 games and won all of them. The truth is, Kramnik only learned how to play chess last year and still gets confused as to how knights move.
We're talking about Boris Kramnik, my 9-yr old cousin, right?
I am very bad at chess, so I guess it must be true!
You are indeed correct in that every two player zero-sum game has a 'solution' and a 'perfect strategy'. Theoretically. Solving chess as a matrix game is, however, not possible in practice due to its extreme size. If you had infinite processing power and memory, it'd be trivial though :)
As all the Lexx fanatics are aware, this seasons final episode involved the complete destruction of an alien vessel 2/3 the size of the earth into a piece of matter the size of a pea. This was done by calculating the exact mass of a higgs-boson particle using a 'portable' super colider.
So perhaps the next finale for lexx should involve building a 1 KG device to 'solve' chess in half a second.
https://www.gnu.org/philosophy/free-sw.html
It is an unwinable game.
This is why it is possible to create a computer that can't be beaten.
Shame on you for not remembering "War Games".
While it may possibly be determined at some point if Chess is an 'unwinable' game, at present time, it is beyond the scope of any programmer to design a chess program (to beat a chess master) that isn't (at least partially) based on known human strategies, that another human can develop a counter strategy to given enough time and assuming they are good enough.
Say I want to put your peices on "BAD" squares. So I check your king early on, to force you to block check with a pawn, then i set up an exchange which puts two more pawns in the way or your bishop. I've just blocked your bishop from attacking giving me a peice advantage. Lets say i move more pawns forward and slowly take away spots your knight can move to.
What you have here is no good moves to choose from, you've lost control of the game, all my peices are on the best squares while all your peices are trapped behind pawns and have poor angles, by using checks, timely exchanges, and etc, if planned right you can easily TRICK a computer into giving you control of the board.
Let a simple exchange of peices can have the end result with my peice on a better square, a simple check can put your pawn on a square i want it to be on, a simple THREAT via my improved position can force you to move your peices to defend against it, i've effectively taken control of the board and you'll spend the entire game reacting to my every move struggling to fight your way out of checkmate, threats of checkmate, and trying to get your peices on decent squares.
What good is your knight if its in the corner of the board because i put your knight there via some exchange which forced your knight to go there.
What good is a bishop if its behind a pawn because i PUT the pawn there when I checked you.
Seems like a wasted move, but that pawn blocking your king was your BEST and ONLY move, and it happened to give me control of the board.
If you use Linux, please help development of Autopac