Test beds for AI
I came across this citation in an article on a new world record set by a computer Go player:
After the victory of IBM's Deep Blue against Garry Kasparov, the game of Go has replaced chess as a test bed for research in artificial intelligence (AI).
Read More →
^ TOP
Posted on 03/06/09 at 11:18:36 by dchaley (David Haley)
1 comment
Thoughts of GGP prover implementations
I always knew that the logical prover I wrote for CS227b back in 2006 was not terribly efficient, but recently I learned just how inefficient it is. While it takes on the order of a second to list legal moves in the first state of chess, a Prolog implementation takes on the order of 100ms. An order of magnitude is a pretty impressive improvement. What follows is an exploration of the inefficiencies of my implementation and my thoughts on improving it.
Read More →
^ TOP
Posted on 03/15/08 at 03:36:15 by dchaley (David Haley)
No comments
Final Round of the GGP Competition
The final round of the general game playing competition is scheduled to begin very shortly. You may watch the games at the same link as yesterday. We're playing some test matches now; if you click on the "matches" link you should be able to see them under the active section.
^ TOP
Posted on 07/23/07 at 11:28:58 by dchaley (David Haley)
1 comment
AAAI General Game Playing Competition
Hello from Vancouver, Canada! In a few minutes, the third iteration of the AAAI General Game Playing competition (or, as Mike puts it, the First Ever Third Annual Competition) is set to begin. We will be running quarter-final matches all day today, followed by semi-finals and finals on Monday. We might be hosting a human-computer competition later in the week, depending on various logistics.
If you'd like to watch, you can do to the tournament page. There you will see a list of active matches (at the time of writing, none) and a list of matches ready to be started. To view a match, you first click on the match key (e.g. "Match.33941089272") and then click on the magnifying glass.
Good luck to our competitors!
^ TOP
Posted on 07/22/07 at 09:57:02 by dchaley (David Haley)
No comments
Tic Tac Toe game network
In the GGP project we represent and communicate games in the Game Description Language, a logical language that very compactly describes games. Due to the semantics and restrictions we have chosen, the class of games we can represent is precisely the class representable by finite state machines. However, the finite state machine for even a simple game like Tic Tac Toe is very large, and would be impractical to store in memory much less transmit to players.
Read More →
^ TOP
Posted on 05/17/07 at 00:02:42 by dchaley (David Haley)
No comments
Game Player Released
The game player I referred to a few weeks ago has been released. Please see the Jocular home page for more information.
^ TOP
Posted on 04/30/07 at 01:46:43 by dchaley (David Haley)
No comments
Game Player Code
We (the Stanford Logic Group) will be releasing a basic program for playing General Games with our framework that I am writing. I have baptized it Jocular, which means "characterized by joking", or "playing" from Latin -- English has put more emphasis on the joking aspect. I will also be working an "official Stanford Logic Group" game player with at least another fellow from the lab, Eric Schkufza (who is also a Master's student). Don't have a name for it yet, though.
Jocular will be the "parent" of Paulatim, the starter code I'll be releasing to the CS 227b class. Paulatim is Latin for "little by little", which I think is fitting given that the code will be distributed piece by piece to the students as they work on their projects.
^ TOP
Posted on 04/08/07 at 01:37:45 by dchaley (David Haley)
No comments
Paper idea for GGP
As I mentioned in my post "Winter quarter", I'm working on a paper for CS 224M: Multi Agent Systems that covers General Game Playing. Although what I'm working on is not multi-agent per se, it involves techniques used in class. Hopefully, it will actually lead to something, in which case I will be able to use it for my GGP work.
Read More →
^ TOP
Posted on 03/11/07 at 01:05:32 by dchaley (David Haley)
No comments
Progress on game player; class competition
We've been pretty busy with our general game player recently. The class is almost over, and on Wednesday there will be a competition pitting all of the game players against each other. The winner is likely to get an A+ plus some kind of mystery prize. So yes, we've been slaving away trying to get our player ready for the competition.
Writing this player has been difficult, to say the least. There are lots and lots of gotchas, but most importantly the program has to be very stable. If it dies in the middle of a game, you still get to participate but all the rest of your moves are made for you very, very stupidly. (The game master finds the first legal move and picks that.) It's not so hard to make your program stable if you don't try to be smart about playing the game, but as soon as you start playing the game "intelligently" you can seriously shoot yourself in the foot.
Read More →
^ TOP
Posted on 06/06/06 at 04:52:20 by dchaley (David Haley)
1 comment
GGP: Introduction
As I mentioned in my previous post, I'm spending almost all of my time this quarter on CS 227b: General Game Playing. So just what is general game playing? Here is the Stanford Logic Group definition:
A General Game Playing System is one that can accept a formal description of an arbitrary game and, without further human interaction, can play the game effectively.
Read More →
^ TOP
Posted on 05/14/06 at 07:47:34 by dchaley (David Haley)
2 comments
|
![[Valid Atom 1.0]](http://blogs.the-haleys.org/media/valid-atom.png "Validate my Atom 1.0 feed"){kind=small}
