Monday, September 6, 2010

AI and chess....an unlikely combination

Brother Zapalla wrote and discussed Artifical Intelligence in class and in his blog post earlier last week. I wrote a short paper on this topic a year or so ago, and thought I would share it with everyone...

In the past 20 years, as the Information Technology realm has developed, technology has grown by leaps and bounds. In 1981, the first IBM personal computer was launched. It had a whopping 4.77 m Hz processor and anywhere from 16-256 KB of RAM. Now, a little over 20 years later, one can buy a computer for a thousandth of the cost of the first IBM computer that is a thousand times more powerful. The more you learn about machines and computer, the more you wonder how far will the power, capabilities, and intelligence of a computer be pushed. In Abu Dhabi, there lies a computer...rather a network of 64 computers, each with 3.06 G Hz processor and with over 512 GB of RAM called Hydra. These computers all work together for one purpose: Checkmate.



Every since the computer's inception, there have been people trying to code strong chess programs to see how “smart” a computer can be. However, these programs coupled with the speed and available memory of the given computers proved to be no match for the human brain. In as early as the 1960's humans would play computers and would win by a landslide. International Chess master David Levy made a wager of 1250 pounds that no computer could beat him in 10 years. He won that bet, as every computer to come his way was met with defeat over the chessboard. However, technology did catch up with Levy, as he lost his first game to a computer in 1989. Since then, the computers have reigned supreme.

Ross reports a number of world-champion Grandmasters who have lost the battle to computers. Most of them, in fact. Britain's Mickey Adams(ranked 7th in the world), Uzebekistan's Rustam Kasimdzhanov, both fell to computer in the same week. What has caused this change in power? Ross states that “First, the best humans now playing are simply not as good as the former. . . Second, the machines are getting smarter. Their programmers are learning how to counter anti-computer strategies.”(link) These programmers are doing what was unthinkable in the 1960's: programming the computer to think like a human.

The human-computer interface and the idea of AI is a newer part of the Information Technology field, but it is a part nonetheless. The human-computer interface has been relevant for a number of years. However, being able to program a computer that deep and specific is one thing, but going through with it and doing it is something else. A matter of ethics comes into play when this in introduced. Is the human brain just a computer? Can it be simulated? Should it be simulated? These are just a few questions that may arise when facing this decision. This is not an intent to persuade one way or the other, but rather to inform of the sides of this controversy. People may say that technology hasn't gone that far yet. However, Mickey Adams said that, “Hydra plays very well indeed. Very often it plays human-style chess, which is strange.” Also, Bulgaria's Veselin Topalov(ranked 3rd in the world), had this to say after losing to new model of Fritz(a popular program like Hydra), “What have you done to Fritz?”


For those of you interested....

1 comment:

  1. It's interesting to see how computers have evolved, and how the trend is such that computers have such advanced algorithms that their abilities look "human-like". I think it's important to remember, however, that just because a computer can compute a very difficult logical problem (like playing chess), the element of humanity which is a combination of logical and illogical thought processes is more difficult to replicate through computer than the purely logical algorithms developed for computers.
    I discovered this while playing Hearts with a group of friends some weeks ago. We'd all played on the computer, and very often won. When we played against each other though, we all found it very difficult, because as humans we would not always make the perfect algorithm choice, and so with that element the game became much more difficult. I think there are other applications of computers where this difference between humans and computers will be very relevant.

    ReplyDelete