It seemed to confirm what Kasparov thought he knew that he’d successfully dragged the game into a position where the number of possibilities was so mind-bogglingly large that Deep Blue couldn’t make a sensible decision. From Kasparov’s end of the table, the delays made it look as if the machine was struggling, churning through more and more calculations. During their infamous six-game match, the machine would occasionally hold off from declaring its move once a calculation had finished, sometimes for several minutes. In order to beat Kasparov, Deep Blue had to understand him not simply as a highly efficient processor of brilliant chess moves, but as a human being.įor a start, the IBM engineers made the brilliant decision to design Deep Blue to appear more uncertain than it was. That symbolic victory, of machine over man, which in many ways marked the start of the algorithmic age, was down to far more than sheer raw computing power. The outcome of the match is well known, but the story behind how Deep Blue secured its win is less widely appreciated. will have a lasting effect on society.But when IBM’s Deep Blue faced Kasparov in the famous match of May 1997, the machine was immune to such tactics. AlphaZero's performance shows that we have to take artificial intelligence seriously, and proves that we are witnessing an enormous change. It also knows other games that it has taught itself. And AlphaZero can do more than just play chess. AlphaZero would play out, if AlphaZero were slowed down to human speed. It would be interesting to see how a game of human vs. The deep network structure is reminiscent of a human brain, and it calculates more slowly: per game, it assesses thousands of times fewer moves than Stockfish. In a sense, AlphaZero is more human than a normal computer.
Grandmaster Daniel King thinks that AlphaZero will change chess theory and the professional game as soon as it is publicly accessible. As AlphaZero began by knowing no chess theory, it had to develop its own theories. Instead, it plays with "real artificial" intelligence. AlphaZero plays very unusually not like a human, but also not like a typical computer. I recommend that anyone who understands anything about chess watch the tenth game against Stockfish. (There are a few valid objections that discuss the use of hardware and unfair competition rules.) It initially made a lot of beginner's mistakes, but after four hours of self-training, it defeated Stockfish. AlphaZero only knew the rules of chess it learned the game by playing against itself for a few hours. Stockfish knows a lot of chess theory, and is familiar with every game ever played. The best programmes, such as Stockfish, reached such a high level years ago that humans stand no chance against them. Since Deep Blue, chess programmes have been continuously refined. But the wolf is really here, and it has a name: AlphaZero.ĪlphaZero, a computer programme from DeepMind, can teach itself skills such as chess. I think that AI research and its advocates have cried wolf too often – meaning that now, people don't tend to see the value of real successes and genuine breakthroughs. Finally, IBM's Watson won the quiz show Jeopardy! – but was it truly down to intelligence, or was it just a case of advanced googling? A few years later, a car from Stanford succeeded in navigating autonomously, but even that was only only in the desert. Critics complained that chess was not a real test of intelligence, and repeated John Searle's Chinese room argument: Deep Blue didn't really understand the game, it simply calculated very quickly. Eventually, researchers didn't even want to talk about "artificial intelligence", preferring instead to study the more modest field of "machine learning".Īt the turn of the millennium came a few successes: IBM's Deep Blue defeated Garry Kasparov at chess.
Even more optimistic were the predictions that computers would be able to perform any and every human task by 1985 – but even today, we're still a long way from that being the case.ĭisillusionment often followed, large-scale research programmes were scrapped, and a series of AI winters set in. It was expected by 1968 it happened around 30 years later. For example, they announced that a computer would become a chess grandmaster. Over the last 60 years, artificial intelligence researchers have continually made futuristic predictions.