The next home game will be Sunday, 20 February at 7pm. Please let me know if you are coming by next Friday (18 February). Congrats to Peter and Colin for coming in the gold at the first game!

Principles of Knowledge Engineering and Reconstruction

The next home game will be Sunday, 20 February at 7pm. Please let me know if you are coming by next Friday (18 February). Congrats to Peter and Colin for coming in the gold at the first game!

1 Comment

We will talk about using evolutionary algorithm to solve poker game including genetic algorithm and evolutionary artificial neutral network. Links to the two primary papers:

Can Opponent Models Aid Poker Player Evolution?

http://inf.brad.ac.uk/~picowlin/PDFs/Conf/CIG2008.pdf

Evolving Nash-optimal poker strategies using evolutionary computation

Other interesting papers include:

An Investigation into Tournament Poker Strategy using Evolutionary Algorithms

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.106.7845&rep=rep1&type=pdf

Computer poker: A review

http://www.cs.auckland.ac.nz/~jrub001/files/CPReviewPreprintAIJ.pdf

Thanks!

Comments Off

Posted In: Classes, Uncategorized

We’ll be discussing two papers: *On the Borel and von Neumann Poker Models* and *Best Response to Tight and Loose Opponents in the Borel and von Neumann Poker Models *with further emphasis on the former*. *You can find links to them below:

http://www.math.ucla.edu/~tom/papers/poker1.pdf

http://math.arizona.edu/~caseyw/poker.pdf

EDIT 2/11/11: In case anyone is getting confused by the missing figures in the second article, I have posted recreations of what I believe the author was trying to show. (Figure 2, Figure 4). The remaining figures are essentially the betting trees from the first article.

Like with the AKQ lecture, we’ll also give you some time to play the games to see what they’re like. We’ll use the following site to help generate random numbers:

http://www.cs.virginia.edu/~mky7b/cs6501poker/rng.html

Additional papers related to the topic are given below that go into the math and derivations a bit more heavily. These won’t be discussed but are provided in case you are interested.

http://www.archive.org/details/theoryofgamesand030098mbp (*Theory of Games and Economic Behavior*, focus on section 19)

http://www.math.ucla.edu/~tom/papers/poker2.pdf (*Uniform(0,1) Two-Person Poker Models*)

http://projecteuclid.org/DPubS?service=UI&version=1.0&verb=Display&handle=euclid.dmj/1077468915 (*A Continuous Poker Game*)

Billings’ dissertation, chapters 1-3, will be presented. These chapters have already been posted by Dave, on this website. The emphasis will be on chapters 2 and 3, since these are more interesting.

Comments Off

The updated schedule is here. Everyone who is officially registered for the class should be listed once (as either a first or second class leader). If you think you are in the class but are not listed here (or you are listed here but don’t think you are in the class), please let me know.

Comments Off

Here are two challenge problems based on the AKQ game. Challenge problems are open until the first good solution is accepted, and may be reopened if a later, better solution is found. Winners of challenge problems receive the undying glory and admiration of myself and their classmates, and have an opportunity to present their solutions to the class. Note that the intent of challenge problems is to encourage you to think creatively about “open” problems; it is possible, though, that some of these problems have been considered by others before and you could find answers on the web. This is not encouraged, but if you do find a published solution, you can still submit it (crediting the source, of course), and although it will not cover you with as much glory as an original solution, it may still be worth presenting.

We determined that the original AKQ game is advantageous to Player 1: when both players follow optimal strategies, player one expects to make a profit. One way to make the game more fair might be to give Player 2 more strategic options. In the “full-street” AKQ game, the rules are:

- Player 1 can bet or check one unit. (as in the original game).
- Player 2 can either (1) fold, (2) call, or (3) bet/raise one unit (adding new strategic option #3).
- If Player 2 bets/raises, Player 1 now has the option to either fold or call.

Describe a solution to this game, explaining the optimal strategies for each player and the expected value for Player 1.

In the original game, all best must be one unit. Consider what happens when this restriction is relaxed, so now Player 1 can bet any amount between 0 (equivalent to checking) and his total stack (n1). Assume the bet has no quantization, and can be any real number between 0 and n1. Then, Player 2 either folds or calls. Note that if the bet exceeds Player 2′s stack, Player 2 can call (“all-in”) and Player 1′s effective bet is the size of Player 2′s stack (n2). As in the original game, each player antes one unit before the action. (What matters, of course, is the ratio between the bets and the antes, not the absolute values.)

How does this impact the optimal strategies for each player and expected value of the game for Player 1?

Comments Off

**This post has been superseded by the Schedule page.**

There is no class this Thursday. The class will be rescheduled later.

The revised presentation schedule is:

- Tuesday, 1 February — Lecture on Game Theory Background
- Thursday, 3 February — Lecture introducing Machine Learning
- Thursday, 10 February — Ben Kreuter, Ben Rodes (Planned Topic: Billings dissertation chapters)
- Tuesday, 15 February — Matthew Yu, Yuchen Zhou (Planned Topic: Borel and von Neumann poker models)
- Thursday, 17 February — Yajun Wang, Weikeng Qin
- Tuesday, 22 February — Xiaoyuang Wang, Colin Braley
- Thursday, 24 February — Peter Chapman, Brielin Brown
- Tuesday, 1 March — Sahnghyun Cha, Hui Shu
- …

Comments Off

Posted In: Classes

Here are the slides from Class 2: [PDF] [PPTX]

The video I tried to play in class is here: *http://www.youtube.com/watch?v=12rNbGf2Wwo*.

Comments Off

©2011 cs6501: PoKER | Powered by WordPress with Easel | Subscribe: RSS | Back to Top ↑

## Recent Comments