Deep Blue beat G. Kasparov in 1997
Вставка
- Опубліковано 12 тра 2007
- Short documentary about computer chess history up to the third millennium and especially about the 1997 chess match between Garry Kasparov World Chess Champion and IBM's computer Deep Blue. The computer won the match 3.5 - 2.5 and Kasparov lost a chess match for the first time in his life.
____________________________________________
Un minidocumentar despre meciul de sah din 1997 dintre computerul Deep Blue si Garry Kasparov, castigat de catre masinarie. - Спорт
PLOT TWIST: the guy moving the pieces for the computer is actually in charge of moving the pieces
and the plot thickens
Twist in plot twist : Kasparov lost on purpose
what if the asian guy moving the pieces is actually kasparov in disguise
That is actually a nice premise for a movie
Plot plot twist: I'm not saying it was aliens... but... it was aliens 💥😂
Kasparov is probably the most iconic chess player of all time. Especially his games against the machine were historic.
chess tictacs I haven't pondered this for a decade, having played chess my whole life, I know that Kasparov vehemently claimed that Deep Blue's programmers were cheating by analyzing how Kasparov was playing, but never come across any relevant evidence to suggest such cheating took place. I really wish to know more about this. Did Kasparov 'choke' when he realized he wasn't going to 'win' the rematch easily? This has also fascinated me. Was the rematch 'rigged?' between intervals so to speak, which Kasparov will go to his grave believing?? Or did the computer genuinely out wit any human's ability to read the game against a computer??? He never stopped talking about the relentless interventions with 'deep blue' between the rounds from multiple chess masters, thus making the match 'fixed' so to speak. There never was a trilogy. So what's the truth? Any big chess fan please illuminate me
@@hungrycrab3297 Do you know why they never payed again? Despite Kasparov wanting to? Thankyou for the insight
@@hungrycrab3297 Do you know why they changed the rules for computer systems like Divine not having intermissions during breaks when playing chessmasters?????
Fischer
i just don't understand why kasparov had to play as black 4 times. it should be 3-3 if the match goes all the way, (best of 6 games).
Reported for botting.
+Kåpla Kvëhla What are you doing here, Kappa Senpai?
Chess equivalent of an aimbot
*Reported for humaning* -Kasparov
@@PaladinswordSaurfang EXACTLY
@@PaladinswordSaurfang
Back here 2 years later : Garry cheated Judit Polgar in 1994 & Vishy in 1995 yet baselessly accused IBM in 1997. What a hypocrite.
a computer can beat a chess master however it cannot experience the joy of winning the match
we can program it to.
Mafon2 PlaySound(Assets.Sounds.Celebrate, PlayMode.Async);
UiManager.Load(Assets.Ui.GameOverPanel).Show();
How are you going to program emotion? You can program a computer to display certain emotions based on outcomes, but that's just displaying a message conveying emotion, not actually having an emotion. Computers aren't sentient.
You can do this by a reward-mechanism. Whenever the AI achieves something they will get that "reward" that's programmed. Maybe some extra processing power or something else that's usefull for the AI. There are multiple theories about this.
@@GijsvanGaalen At best, that simulates emotion, and not emulate it.
The first documented RNG ragequit in history?
+Bob Shanely rng... in chess? if it's a joke, i don't get it.
+busTedOaS "For his book, Silver interviewed Murray Campbell, one of the three IBM
computer scientists who designed Deep Blue, and Murray told him that the
machine was unable to select a move and simply picked one at random. " www.wired.com/2012/09/deep-blue-computer-bug/
+Bob Shanely Thanks. That is actually very interesting. The best move was random... it makes me think whether it is necessary to have intent in order to qualify as intelligent. Or maybe it was even the best move to confuse Kasparov with that random ass move... just to get into his head. although I doubt that was Deep Blue's strategy.
@@BobShanely No, there is a system which determines the advantage and disadvantages of a move with a rating. It would simply take the one with the highest rating. Also it wasn't just brute force like they said in the video. The programmers were smarter than that.
The fact that he managed to beat the computer more than once is an amazing feat in itself. I would have thought that a computer that can process millions of moves every second would be unbeatable.
It is incredible.
And, in fact. No longer even possible.
AlphaGo can still potentially be beaten at Go, but even that is starting to look impossible.
Luckily, we will have Mao and Calvinball for a long time yet.
Well theoretically it is possible, just unlikely
Nathan Schubert and now alphazero beats stockfish8 (which wouldve stomped deep blue) 100 - 0
I personally think they handed him the first game so the computer can learn his moves
@@MagnumSkyWolf There was a bug in the program in the first game, they fixed it from game 2 on.
If I remember correctly: one of Kasparov's complaints was that humans interacted with the computer during the game, so in a way Kasparov was playing against more than just the machine. IBM said that that was simply the way the machine worked - it required the intervention of programmers - but I sympathize with Kasparov's point of view here.
Well now computers can beat humans on their own
Damn bruh the fact that its been 14 years since you commented and you were using yt at that time this gives me a nostalgia of something i never experienced
And now there is AlphaGo.
And AlphaZero
Leela Zero, another self-teaching AI, has recently showed up!
AlphaGo vs AphaZero??
AlphaGo plays Go, AlphaZero plays chess. They're different games.
AlphaZero plays GO and SHOGI besides CHESS. but I think AlphaZero will win against AlphaGo(the one who beat lee sedol) but not against the latest version of AlphaGo.
Kasparov could have won or even tied the Match if he hadnt be intimidated by enourmous procesing power of IBM machine. For example in game 2 he resigned wrongly cause he didnt notice could draw the game. Another mistake was to try use anti computer strategy, by playing bad moves in openings to confuse the machine. Although calculate milions of moves per second, Deep Blue was not perfect and had several bugs in its positional play, wich scientists tried to fix during the match. If Kasparov had played normaly as would do against a human oponent, he could have won. Deep blue defeated him psychologicaly.
+Miguel Eduardo RESPEKT! "Deep blue defeated him psychologicaly." :)
another person who is better at chess than a world-professional chess player?
still a win nevertheless
and thats why cpu's are better, no emotion, only strategies.
@Justin Y. Considering that Miguel Eduardo's point's validity hinges on him being better at chess than Gary Kasparov was...
As amazing as it was in 1997, to think that a program in 2017 like AlphaZero could totally crush Deep Blue (using AI instead of brute force) is absolutely mind boggling.
i just don't understand why kasparov had to play as black 4 times. it should be 3-3 if the match goes all the way, (best of 6 games).
Stockfish uses brute force and is brilliant
AI is brute force.
2024 and Stockfish could crush 2017 AlphaZero
No one could play better than Kasparov, and his natural reaction after losing is just priceless.
i just don't understand why kasparov had to play as black 4 times. it should be 3-3 if the match goes all the way, (best of 6 games).
260 processors versus one man... That sounds fair.
Funny, a few years ago it would sound unfair to beat a machine using our "great" human brain.
How is that comparable? I could also say 260 processors against 86 billions neurons. That just means nothing.
Djorgal lol it has all combinations isntalled. It can analyze milions of moves in seconds and one man cant handle that much info. Yes one brain can have alot of memory, but analyzing all moves fast or think of combinations as a computer is not even comparable lol. One computer would even beat all gms at same time. If you are so smart try to challenge not computer, challenge something easier like calculator. Multiple 25687x26741 and start thinking when you click on equal button. I bet you will not calculate it fast to beat the calculator right? Same here, even standing a chance against this super computer is a huge deal!
And take the same calculator, show it a picture, and ask it if that is a picture of a bird.
It won't be able to answer. Humans are good at doing some things when computers are better at doing others.
To compare the number of processors of a computer to the number of brains of a human just means nothing, nor does it mean anything to compare numbers of transistors to numbers of neurons. Those are not the same thing and that doesn't make for a sensible comparison.
A human chess player can realize what a fork is. If a knight attack two of his pieces at once, the computer needs to see at least one move ahead to conclude that it loses one.
The grandmaster doesn't need to make the calculation at all because he knows what the position means.
Yes the computer can analyse millions of moves per second but the chess player can actually devise a strategy.
Djorgal Your first sentence answered your question. "It wont say it is bird" but it will beat any human or at least draw but never lose in calculation, because it is programmed to calculate. If it was programmed to identify picture, trust me it would beat your ass anytime. Deep blue is programmed to beat ppl in chess and it is doing its job. If it was programmed to do everything human can do it will do, but never go beyond the limit of that. It can defintly not do better than that, neither can kasparov, because it is the limit be able to analaze all the moves possible. What you talk about is artificial intelligence, which is an intelligence that has the ability to think, manipulate, make choices independatly and creating some type of logic to understand things. There are no or wont be any machines that are artificial intelligence, but there will tons of machines that wil be prgrammed to do a particular job, far beyond the posibilites of human brain abilities.
basically a group of intellectuals beat 1 genius
+Dan0101010101010 Group of genius beat 1 genius.
A trained genius beaten by 1 box of silicon
+Dan0101010101010 Imagine what a group of geniuses could do to an intellectual? That would be like upgrading you to +Dan1111111111111 = Dan Powers
Yes and no. After they taught the computer (Programmed), the computer does all the work (Brute Force technique).
Basically a box beat 1 genius.
"Scientists finally built the first artificially intelligent machine. The first question they asked it was 'is there a god?' It's response? 'There is now.'"
:)
Heh. More sensational than witty, really. True AI doesn't immediately mean a Skynet.
Nor would any self-respecting scientist ask it such a ludicrous question.
Banzay27 According to Steve Hawking, it does.
Banzay27 You don't get the reference.
logandh2 It's response?
"I'm not sure, but I'm really good at chess."
I feel sorry for Kasparov. The chess programms were weak until 1997. But after 2005 that's impossible to beat a chess computer.
This video is almost as old as UA-cam itself :0! I still remember watching it in 2007.
Honestly love to see Ken freaking Thompson in the crowd. One of the forefathers of computer automated chess.
And father of UNIX, the operating system that changed the world, and many other awesome things he contributed to (UTF-8, GoLang, Plan 9, and many other).
Gary Kasparov thought that he could out smart the computer, Deep Blue prevailed because it had multiple ways to solve the same problem. Kasparov was surprised by its mathematical ability to play and not reveal its strategy.
one of the best short documentaries ever
Still I like Gary Kaspaov''s playing style, and it is a matter of sacrifice
The ending is profound- it looks to me a late 90s documentary, mentioning that a computer that could think itself is far away- well, the age of AI is here!
Remember when Kubrick took IBM, and translated its letters back in the alphabet to make HAL? And then HAL beat a man at chess in the movie in 1968?
I was always glad to see that IBM took it upon themselves to make sure things came full circle.
The computer could study many, if not all, of Kasparov's games.
The matches Deep Blue had played with other GM's were private, so Kasparov knew nothing about his opponent, while his opponent, if human, could write an essay on his play styles and how they evolved over the years.
Totally fair.
Nice video ,actually epic
I taught myself how to play chess at 9 years old. And I’ve sucked ever since.
This match was played when i was 9 years old but i remember like yesterday i think garry kasporov real legend of all time in the world 👍
Thanks for the video. Do you have any video links about games between grandchesmaster and chess machines? Thanks.
This gives whole new level to saying " I'm not a Robot "
Cuando un día que no veremos nos gobierne una máquina como Deep Blue, será un héroe como Kasparov el que nos libere, alguien que no sea perfecto, pero tenga pasión.
.
its weird how this is the only video with archived footage of the event actually happening
Hi Kow Im Cobra
@@idisplaypace2411 hi i am gonna go play blitz
@@BlitzWizard94 why bro
@@idisplaypace2411 cause i am the BlitzWizard
Yeah I wish there were surviving live footage of the match. I still don’t have any idea about how much time deep blue takes to make a critical move, reactions of Kasparov in tense postions etc. Wish I could spectate the match 26 years after it was conducted
In 1997, Skynet goes online
I was reading about this match in a book, it is astonishing to think that what once in 1997 was shockingly state of the art technology nowadays is considered a milestone in the path to the future of AI, thus the future of the entire human race
•1996 Deep Blue Vs G, Kasparov 2775 4 2 Human Wins
•1997 Deep Blue Vs G, Kasparov 2795 2.5 3.5 Computer Wins
••2007 Krempov vs Rybka 3150. 6 game. Krempov draws 2.5 2.5=Unafisual Worlds Record. •2007 Krempov Vs Deep Junior 10. Rated 2900. 1 Game Match. Score a draw.
•2014 Krempov Vs Stockfish. Rated 3250 In a 6 game match krempov wins 4 to 2
•2015 Krempov vs Stockfish5 rated 3280. In a 3 game match. Krempov wins. Krempov's now chess rated is 3285
In the future they will roll this video on schools to show us the beginning of the fall of human civilization way before the War agains AI.
Sip
And how other humans like these programmers were so eager for the obsolescence of humanity.
He's an amazing brain ! To beat a specially made computer with all the possible moves in its memory is quite something,
What Is crazy Is that Kasparov pulled draws and stood a chance that has million options every second, if you really let that sink in you can imagine how crazy smart his brain is..
So impressive at the time but now he or Magnus stood 0% chance of winning.
whos here after alphago
+Vlad Novetschi i am. and i am a go player
me 2 m8
Wanna see Magnus Carlsen vs deep blue?
its not possible since after Kasparovs loss they destroyed the Deep Blue
port the code over to a computer today
At this point, Stockfish beats Magnus easily.
@@SimsHacks why can you tell us?
This is one of those moments in history that warrent a whole blockbuster movie budget
In 2022, this is a video which I can call a classic epic 🤩
I'm programming a chess engine myself right now, have a bet with a friend of mine who's a tournament player and No. 5 on our state's leaderboard, wish me luck!
What happened?
@@Wolfboy950 I won! The engine beat him in 4 out of 6 matches.
@Sztanyi you are completly right, I have been playing chess for more than 12 years and I beat some chess programs like "rival chess" "master chess" and another ones, but I try to beat at least the "Arasan" and I think it´s so hard, so in my opinion Kasparov is a hero, he is the best of all.
So iconic and historic, I wonder what museum Deep Blue is preserved in?
If you watch the documentary Game Over (it's on youtube), you'll see the kinds of secrecy surrounding their computer, plus the controversy over the one move Kasparov complained about which tipped the scales for the whole match.
@soccom8341576 Yea i know :) it was a wide variety of people too, from engineers to other chess grand masters and full gameplays from a bunch of chess books written by other chess grand masters.
That fact just makes me even more amazed at what Kasparov was able to do.
Not to mention Deep Blue was constantly getting tweeked and upgraded to better counter Kasparov's playing style. To me it is truely amazing
20 years ago today.
Wonder how did i end up here ? Strange... I was searching for an album
@JOTTABYTE Surrendering in a game in which you overlook a draw opportunity is still a loss. It was in game two that Deep Blue 2.0 made an unpredictable move that completely frazzled Kasparov. In game six, Kasparov saw that he could no longer out-think Deep Blue, so he gave up.
Kasparov proved himself human: giving up out of exhaustion, acknowledging a loss before it becomes apparent.
"Nevertheless, the machine they had built did not play chess by thinking in the same way a human does. A machine that can think remains the dream, and it's still many years, and quite a few startling breakthroughs away."
And so the creation of AlphaGo began.
The Famous Equation :
kasparov + deep blue = me
The thing is that personal computers use general-purpose processors, which are designed and programmed to perform fairly well in any aspect of usage (playing video games,handling multimedia,exchanging all kind of files over a network, etc). You could call them Jack-of-All-Master-of-None processors.
On the other hand, the special-purpose processors are designed to perform at maximum possible level, but only in a very small area (often limited to just a single application/program).
Back then you could accuse a computer of sing a human brain today you could accuse a human of using a computer brain
The documentary implies that Deep Blue uses "brute force" to win. Indeed, it looks at millions of moves, but it intelligently decides which parts of the future move trees to explore.
Well, "intelligently". There's just well thought algorithms and logic, which one can say are designed by intelligent people. It could be said to be intelligence in the machine if the machine figured that out itself on the fly, but it did not. People focused on this very problem developed and tuned the machine and its software to solve this very specific problem. For any other type of problem, the intelligence of the machine is 0, until someone again "teaches" it.
200.000.000 million possible moves on every second.
200.000.000 or 200 million ;)
I don't think any human have a chance against a thing that can think about 2 000 000 moves in a second.
I like the hint at watson in the end
Kasparov was asked in an interview after beating Deep Thought in 1989... "How do you feel that someday, you do lose to a computer?"... Kasparov replied "Single game of the match"... And that loss happened in a single game of the match with Deep Blue in 1997....
This is a mile stone in the series of events that lead to an ai becoming self aware. skynet = game over
. . . . . . And John Connor is riding his dirk bike around
+ForTheHonor So many paradoxes.
No, not at all. Deep Blue is just running an endless loop of probing moves and chooses the best one. No different than running a brute force password breaker.
@Eustake whats the background song called ? pls any1 answer
@ollecarlsson That's still pretty far away. The `problem` is that there are billions and billions and billions of possible outcomes of the game, so the programmers have succeeded till present day to simply enlarge the computer's opening database (which shows which moves are the best at the start of a game). After 15-20 moves, the computer must simply use `brute force` to calculate outcomes and choose which moves give it the advantage. This happens due to the enormous number of possible positions
Soon computers will be better than us at everything.
not at everything at the same time tho :).. Our brain is still x100 times better than any computer this size
if that happen we have created them so we are better xd
Alexis Vásquez honestty i dont believe that we can create something better than us considering that we are improving day by day
WhoAreYou Technology is improving exponentially faster.
Alexis Vásquez In what sense and why?
But the human brain just doesn't use stored patterns. We take risks, contextual risks which computers can't make.
Like it's stated in the video the machine basically tried all possible combinations and picked the best one. A human can do the same but we just don't have enough time to assess every possible outcome.
it cant do that. there are more possible moves in chess than there are atoms in the universe. it needed a opening database and an end game table but in the mid game it will use its processing power to calculate many moves ahead
john li Huh? More chess moves than there are atoms in this universe? LOL.
How many atoms do you know? What about those we don't know about? How far around the universe have you travelled, away from our tiny little speck of dust we call home?
I believe the number of chess moves is a lot but finite, therefore calculable. Since we know the maximum number of chess pieces that can be on the board at once, we can calculate all the possible moves... using simulation of course.
Also, the fact that the A.I is domain specific (i.e. its sole purpose is to play chess only), it will only get better in that dimension while still being bound by the rules of that domain, which is chess rules. It will never think outside of it.
Human Being there are 10^120 possible different chess games the computer cant calculate them all and it doesnt have to.
Now we want a same video on AlphaZero
All this really proved is that it takes a team of super engineers years to build a computer after decades of research with a library of human games played before to match the ability of 1 man doing it in his head.
no chess player can beat an ai ever now
It's kind of scary... It's turning into this man vs machine thing... Machines are becoming more complex and one day might take over.
The succession of which? We destroy the earth...
Yaa i am afraid too. My vacuum Cleaner tried to kill me today LOL
Take over what, and what for?
Djorgal
They can do almost everything better than humans... it's insane.
***** No it didn't knew the moves in advance, its memory would never have been enough.
It was calculating its next move using an evaluation fonction. It does evaluate each move.
they used a brute force algorithm... is that right? it seems like IBM would have smarter programmers.if they won using brute force and super fast calculations, imagine what they could do using a better algorithm than brute force.
Sergio Fernandez I read an interesting article about a theory relating to that sentiment.
It goes like this, once we achieve a certain goal in creating an A.I(although the article was about Neural networks), the idea ceases to become magical. That's when we say "Well, nothing revolutionary here, it's just good old brute force", therefore demeaning the A.I.
It's funny because ultimately if you look at every form of A.I out there, that's always the case.
The only way I think to create an A.I on par with a human is by way of a collective hive mind. Something like SkyNet, where each unit has a very specific task of collecting information (visual, auditory, olfactory, etc), then that can be linked and associated with everything else. The information would then be used by other/same units for decision making.
@Sztanyi He also managed to win games against Junior or X3D Fritz, years later. However those computers seem inferior to Deep Blue, considering the number of positions calculated per second.
@ace942 They are trying some interesting stuff such as: giving the human a two-move advantage in the beginning, taking away a pawn from the computer, giving the human the first move in each game of the match etc.
No... Kasparov was not defeated. I mean come on a teraflop super computer vs a human brain, I mean seriously. Is like saying a ferrari run faster than the worst faster runner. Fail.
ikr
Current chess apps AI are leagues beyond Deep Blue and *any* human player.
Yes I fully agree. However you're a total fool if you think analyzing millions of gamestates per second is consciously obtainable.
+Super Dude
@Ecrilon you are right on all except for the last one which is the possible moves. Look up Scientific American magazine in the october issue 2009 and you will see an article that states explicitly chess has 10^20 moves and go has 10^80 moves, which has as many moves as there are atoms in the entire universe in the game of go.
@CamButler True. Though Deep Blue was still considered a supercomputer.
The algorithms will have improved too.
Gary Kasparov is THE man who incite me to play chess
where can I find the complete documentary
What about the pawn sacrifice issue during Game 2? Why wasn't that addressed?
Kasparov- you can't defeat me
Scientists- but he can
Deep blue emerges.
Chess has seen much more suspense before and since this match.
actually the reason why Deep Blue was "faster" was because it relied completely on brute force calculation. While chess engines like Houdini or Rybka have extensive evaluation functions making the calculations a lot slower but at the same time making the chess engine much stronger. If a modern PC with a core i7 processor used a chess engine that with a primitive evaluation function, it would calculate MUCH more than 200 million moves per second
What is the name of the song playing in the background?
what is the musical piece during the ending credits?
where is this doc from?
@Eustake That seems again unfair to the computer. The human does not begin processing from the beginning. The human draws on his/her own experiences to shape the opening. Also, in modern computer chess, you have access to the computer's opening book, which is restricted and does not run as deep as any master's opening repertoire.
what’s scarier is that chess computers to this day can still be improved. if google holds true then there’s 10 to the power of 111 possible chess combinations, for those who don’t do maths that’s 10x10, done 111 times. theoretically the strongest AI possible would be able to calculate every move possible from start to finish
Whats the name of the first background song?
So this is where it all began
To be honest, I bet no man can beat inexpensive computer loaded with Arena Chess under Houdini engine.
I may be looking at the problem at a wrong angle, but lets understand that Deep Blue had more than one "Nodes," or processing units. We can assume that the program would've utilized this, and would've probably tasked each processor to think of a possible move (in parallel), since it's computing via Brute Force, of course. Deep Blue had 256 VSLI relatively slow processors, compared to 8 fast processors.
Deep Blue had an OS too, but even if they both ran the same software, DB still wins...
I remember that match. They tweeked the program during the tounament. He beat the program as it was when he was challenged.
Fischer would beat Deep Blue 10-0 and then never give it a rematch.
What a true lee great way to work your mind out .
Music name at the end ,please???
hello, where i can dowload the deep blue software?...troll-a-bit LOL
So true! It will go on to end in Stale Mates.
in each match there is a lot of secrecy. Deep Blue team was not asking Kasparov's team how they wer preparing for the match and what openings he is going to play
Deep blue has been relentless, it has followed Gary till this day.
It wasn't easy as evaluate millions of games, Deep Blue had abilities to predict moves and planning strategies. It was Artificial Intelligence according to the definition of AI
@Gihipoxu Technology improves rapidly. 256 processors don't mean anything. It would take only a moderately powerful computer to beat Deep Blue's search rate (though more than the average desktop). Using improved algorithms however modern chess software searches farther ahead and is definitely capable of being better.
@MrJuno6 They were only allowed to tweak its programming between game matches, not during. The actual AI was the one choosing the moves, the tweaks were only to the AI between games, which isn't really cheating at all.
hey IBM, show the logs of deep blue (Karpov?) acts!!
It took far more sophisticated computers to realize that consciousness infinitely surpasses any finite machine.
What doesn´t change at all, is Human´s Ego. I heard yesterday an old Kasparov´s interview; in which he was asked; what does mean for him "chess". His response was: "crushing and crushing the enemy". The problem with those people, is that once they got "crushed", they´re unable to accept it. I never liked this guy Kasparov. I never liked his personality. After losing, he acted like a 5 year old child and complained with his mom (Im not making fun of it.It really happened that way). I sincerely congratulate the IBM team, they did a great job.