In 1997, Deeper Blue, the second version of IBM’s software, won a match in 6 matches (3.5 to 2.5) against the Russian champion. Since artificial intelligence has made such progress, the intricacies of Japanese shogi and the game of go have been greatly improved in 2017.
Forgetting “ Pawns are the soul of chess». Perhaps because since the Deeper Blue computer’s victory over chess champion Garry Kasparov in 1997, the machine has continued to demonstrate its ability to outpace humans on increasingly complex tasks, without being fully competitive in terms of adaptability and versatility.
Deeper Blue’s victory was an event”Culturally amazing“, But “Technologically speaking, it was just a hit“, explains to AFP Philippe Roulet, doctor of artificial intelligence. Basically, the success of Deeper Blue is just a victory for “brute forceExplains the co-founder of Artefact, a consulting and technology development company in the field with 800 employees.
At the moment, the machine wins thanks to its enormous computing power, which allows it to determine all the possibilities of game development very far in the game, and deduce the best piece to play. In fact, the real revolution at this time is hardly emerging from the labs: it is the revolution of machine (or machine) learning, and artificial neural networks, which have made tremendous advances in artificial intelligence over the past decade.
While Deep Blue knew how to play chess thanks to an entire structure of logical rules instilled by humans, new machines are crafting their own rules, in a learning period as they absorb mountains of data. “We move from compulsory programming to programming by learning”, explains Philippe Roulette. The effect is amazing in games. In 2017, the AlphaGo algorithm of DeepMind, based on machine learning, beat the best player in the world, Ke Jie.
before AlphaGo,”There has been speculation about the handicap that must be given to God in order for man to be on an equal footing with himCedric Villani, deputy mathematician, and author of a founding report on artificial intelligence (AI) said recently in 2018.And then we see the overwhelming level of the human beingby AlphaGo,”We understood that man was less good than he thought. Some moves in the algorithm that we initially thought were beginner mistakes turned out to be great moves‘, he completed.
Today, machines are even capable of defeating humans in games with uncertain universes, such as poker or bridge, as French startup NukkAI recently demonstrated. These giant strides completely bypass the gaming world, as the device no longer has much to prove. In recent years, artificial intelligence has madeA very impressive progress that surprised me myself‘,” explains Yan Lecon, head of AI research at Meta/Facebook, and one of the founding fathers of modern AI.
“Today we are able“to let the machine”Translate any language to any language with a set of 200 languages” where “To have a single neural network that includes a hundred languages“But the machines still hit obstacles,” he explains.Not because we can have a funny dialogue with GPT3“, Elon Musk’s amazing OpenAI startup text generator, this”You will be able to help us in everyday life‘, was the nuance.
What this virtual assistant design lacks – and no doubt also for true autonomous car manufacturing – is access.”To a general method of self-learning“We would put the computer in front of 200 hours of video, and from that, it would arrive at a form of ‘understanding the world,’ common sense. This will then allow him to access his learning capabilities.”Closer to what we see in animals and humans“, he explains again. The researcher is convinced that machines will one day own.”A universal ability to learn, capable of learning anything that humans learn, in most cases with higher abilities“.
“But when will this happen? The answer is not clearHe concludes with caution…
Kasparov’s historic defeat, blow by blow
Deeper-Blue -Kasparov, New York 11 May 1997, Caro-Kann defense Substitute Karpov
1.e4 c6 2.d4 d5 3.Nc3 dxe4 4.Nxe4 Nd7 5.Ng5 Ngf6 6.Bd3 e6 7.N1f3 h6 8.Nxe6!?! (a) Qe7 9.OO fxe6 10.Bg6 + Kd8 11.Bf4 b5 12.a4 Bb7 13.Re1 Nd5 14.Bg3 Kc8 15.axb5 cxb5 16.Qd3 Bc6 17.Bf5 exf5 18.Rxe7 Bxe7 19.c4 Kasparov resigns 1 -0
a) a decision that, at the time, is unbelievable or even impossible to make for”MachineShe, IBM’s super-calculator, embarks on a design, a speculative sacrifice, hitherto the exclusive privilege of human thought. An annoyed Kasparov will claim, after the match, in blunt terms, that it was the great educators and computer scientists who drove the program’s choices at the last minute.. .
Right hand under his chin, his eyes darkening, Garry Kasparov stared at the chessboard one last time, annoyed, before suddenly leaving the table. It’s Thunder Bombing: The Chess King has been beaten by a computer. On May 11, 1997, the World Champion Machine was knocked down for the first time after an organization match. This date will mark the history of the discipline and highlight the amazing potential of artificial intelligence.
L “Baku ghoulThe 34-year-old then a master of the world chessboard since 1985, lost after six matches, in a final score of 3.5 to 2.5, to the IBM-designed Deeper Blue supercomputer.
In front of cameras from all over the world, who came to shoot the show in New York, this setback has a taste of humiliation for the stormy Kasparov. Did he not assert that he would withstand machines at least until the dawn of the next millennium?
Un an plus tôt, le Russe avait mis à terre Deep Blue sur le score de 4 à 2. Mais ses concepteurs n’ont eu de cese d’améliorer le monstre de 1,4 tonne, capable désormais de calculer 200 millions de positions par the second. The hero was shaken by his defeat, but nevertheless refused to accept the superiority of the machine. “The computer hasn’t installed anything yetHe said during the post-match press conference, AFP reported. “Man, the best player in the world, cracked under pressureHe explains, speaking for himself,But we can beat the computer, it has a lot of weaknesses“.
Far from comforting the loser with the $400,000 promised, he also criticizes IBM, which did not give him access to previous games the machine had played, and which was able to analyze all of its own. He even became accused, suggesting that humans aided the computer during the match and regretted not doing so.”You set certain conditions“Till the game be”honest“.
Other world chess players, who have examined the live confrontation, also refuse to consider the defeat of the great leader a turning point. In response to a question from the press, they indicated a series of bad choices for the Russian hero. For some, his intense need to understand the blows of the machine, rather than focus on winning, was fatal to him.
Computer opportunity or the hand of a man?
Years later, the book would pass on a secret from the developer of Deep Blue: a computer malfunction that would have turned the game around. Unable to choose between several moves, the machine could have randomly played during the game, destabilizing Kasparov for the rest of the confrontation.
Either way, Deeper Blue’s win made IBM happy to see interest in its computer prowess.
“All of this has nothing to do with the human vs. machine battle but with how we use technology to solve complex problems.‘, excited after the match IBM Project Leader Chung-Jen Tan. To pay tribute to what AI will bring in many areas, from financial analysis to the study of natural, atmospheric or seismic phenomena.
Kasparov, eager for revenge, would equalize twice in 2003 against PCs. But time will calm his wounded ego in the end. After the 1997 defeat,Nowadays, laptops come with a high-quality graphics card“, he stated in an interview with the Swiss newspaper Le Temps in 2019.”But with a little hindsight“She seemed to defeat him.”victory for mankind‘Because she was warned’The breadth of activities that technology can assistFrom now on, what worries the chess legend is the digital giants’ excesses in individual liberties:We want them to be responsibleIn an interview with AFP in November 2021, he urged.