what constitutes AI?

Code, algorithms, languages, construction...
Post Reply
stackOVERFLOW
Posts: 5
Joined: Wed Dec 11, 2013 11:32 pm
Real Name: Stack

what constitutes AI?

Post by stackOVERFLOW » Sat Dec 14, 2013 3:17 am

The most exciting game for me, among many others I have programmed, happens to be one I casually participated in for 2 months,
the Google AI challenge 2011 - Ants. It seems this sentiment is shared with many other participants as well, and it not just the number
of participants, or the short time one has to produce a bot, that made it so exciting. It has some elements carefully designed to avoid
traditional approaches in game tree search, such as using brute force alpha-beta. E.g. The time limit for making a move was limited to 500ms
that basically forces you to rely on other means. The game rules are very simple but lead to a complex optimization problem between different
competing factors. Go has also simple rules and becomes complex, which is probably why it is considered better than chess as a test bed for AI.
However one would wonder with the success of MCTS with pruned trees, it may be going the same path as Chess, only now the problem is bigger!
Ants OTOH will not allow for brute force search because of incomplete information and other factors. These are my opinions. A summary of
another person's opinion that led me to make this post can be found here. What are your thoughts

stackOVERFLOW
Posts: 5
Joined: Wed Dec 11, 2013 11:32 pm
Real Name: Stack

Re: what constitutes AI?

Post by stackOVERFLOW » Tue Dec 17, 2013 4:31 pm

A BBC article posted today talks about it. They call the success in chess a cheap trick!

User avatar
Chris Whittington
Posts: 437
Joined: Wed Jun 09, 2010 6:25 pm

Re: what constitutes AI?

Post by Chris Whittington » Sun Dec 22, 2013 1:17 pm

stackOVERFLOW wrote:A BBC article posted today talks about it. They call the success in chess a cheap trick!
exactly so

hyatt
Posts: 1242
Joined: Thu Jun 10, 2010 2:13 am
Real Name: Bob Hyatt (Robert M. Hyatt)
Location: University of Alabama at Birmingham
Contact:

Re: what constitutes AI?

Post by hyatt » Sun Dec 22, 2013 5:41 pm

Or, perhaps, the underlying principle that says "playing chess is an intelligent act" is simply flawed. Personally I believe what we do today is just a different form of "intelligence" assuming we are going to call things a human does "intelligent". Humans have no chance in checkers today. And very little chance in chess.

This has been the favorite counterpoint for the anti-AI folks, for years.

"playing chess is an intelligent act".

computers reach the point where they can beat all humans.

"playing chess is not an intelligent act."

Seems more like "I won't ever admit computers show intelligence. Anything they can do is by definition NOT intelligent." Where it used to be "anything we do is intelligent."

Even trying to pin down a definition of "intelligence" is difficult. Is an ant intelligent? Seems that the current definition has become "anything the computer is not."

User avatar
Chris Whittington
Posts: 437
Joined: Wed Jun 09, 2010 6:25 pm

Re: what constitutes AI?

Post by Chris Whittington » Sun Dec 22, 2013 9:43 pm

hyatt wrote:Or, perhaps, the underlying principle that says "playing chess is an intelligent act" is simply flawed. Personally I believe what we do today is just a different form of "intelligence" assuming we are going to call things a human does "intelligent". Humans have no chance in checkers today. And very little chance in chess.

This has been the favorite counterpoint for the anti-AI folks, for years.

"playing chess is an intelligent act".

computers reach the point where they can beat all humans.

"playing chess is not an intelligent act."

Seems more like "I won't ever admit computers show intelligence. Anything they can do is by definition NOT intelligent." Where it used to be "anything we do is intelligent."

Even trying to pin down a definition of "intelligence" is difficult. Is an ant intelligent? Seems that the current definition has become "anything the computer is not."
Clearly, you didn't even bother to read the linked article before mouthing off. Your straw man constructions above have nothing at all to do with the points raised within. There's absolutely no point in discusising with you when you've not read (same as the Zach document), and just assume the content based on prejudgements and cliche. Sorry about that, but the effect is you just act as a blocker to any different or new ideas. Not interested.

hyatt
Posts: 1242
Joined: Thu Jun 10, 2010 2:13 am
Real Name: Bob Hyatt (Robert M. Hyatt)
Location: University of Alabama at Birmingham
Contact:

Re: what constitutes AI?

Post by hyatt » Mon Dec 23, 2013 5:24 pm

did YOU bother reading the article? Here is one quote:

By Hofstadter’s standards, iconic computational achievements like beating the world’s best players at chess or Jeopardy are rendered trivial by the “trickery” involved: by the fact that the winning computer has done little more than weigh the relative benefits of several billion possibilities, without at any point knowing anything about the nature of the game being played.

"trickery" -> "not knowing anything". As I said. Turns out searching billions of nodes DOES know quite a lot.

Try again. You always do. But you never get it right...

Interesting that humans can't beat computers, yet computers "don't know anything about the nature of the game being played." Just more whining. Humans are not quite as "intelligent" as they seem to think they are. Computer chess has served as one great example of that.

Post Reply