In May 1997, IBM’s Deep Blue Supercomputer won a fascinating match with the reigning World Chess Champion, Garry Kasparov. In January 2011, IBM’s Watson handily defeated the two Jeopardy champions during a demonstration round. In February, Watson will compete against Jeopardy champions Ken Jennings and Brad Rutter over three nights. This is another progress in the area of Artificial Intelligence (AI) toward the ultimate test, the Turing test. According to the Turing test, a machine is considered intelligent if a person that chat with a human and a machine cannot define who is who. To achieve it, the machine needs to understand natural language including nuances. This is the challenge that Watson needs to handle as part of the Jeopardy game, and it has to do it fast.
In a different research lab at IBM, researchers developed an automated agent (or intelligent program) that can perform trading in a model similar to the stock exchange market. As shown in that research, the agents outperformed the humans and showed higher profits. While each of those examples enabled machines to be more intelligent, we are still far away from a human intelligent robot.
Gadgeteer Comment Policy - Please read before commenting
There’s a book about all of this! Check out Stephen Baker’s FINAL JEOPARDY: Man vs. Machine and the Quest to Know Everything. It’ll be in stores Feb 17 – but you can get an eBook now, online. Check out this piece on AllThingsD about it: http://tinyurl.com/4ouh4je.
As a software developer with expertise in A.I. and a regular chess club player (FIDE rating 2200), I would like to let you know the following :
The match G.Kasparov vs Deep Blue has been widely considered to be fixed within the chess community, and a cheap advertisement for IBM (I could quote some of the strong indications that lead to this conclusion, but I guess only insiders would understand, and probably they already know what I’m talking about)
To the public, this appeared like a sequence of interesting and fascinating games. But I doubt that it proves anything about the man vs machine battle on the chess battlefield.
Writing a year ago in the New York Review of Books, Kasparov said that the golden age of man-machine battles in chess was between 1994 and 2004. Before that decade, humans dominated. Afterwards, machines. He said that interesting matches featured humans using chess-playing machines. I think this is also the model for Watson, the Jeopardy computer. It’s future is working alongside humans. Its solo turn on a quiz show, while interesting, is a contrivance.