EAAI NSG Challenge: Parameterized Poker Squares

• 5/27/15: Announcements:
  • This is the latest distribution of the base code (documentation) with only a minor bug fix to PokerSquares.java main method demo code.  If teams would like to develop in a language other than Java and the faculty mentor is willing to provide a Java interface to such code, I will try to accommodate.
  • Deadlines have been finalized:
    • August 15, 2015: Parameterized Poker Squares players (one per team) are due via email to tneller@gettysburg.edu  (Subject: Parameterized Poker Squares Challenge Submission).  These must meet base code interface specifications.
    • September 15, 2015: Optional research papers on such players by student-faculty teams are due for submission to the EAAI symposium via the AAAI paper submission system. (More details are to come.)
  • If you've not already done so, please register your team.  Your student team members would have recently received these announcements if I have their names and contact information.
  • Early testing:  As an informal means of early evaluation, teams are invited use EarlyPPSTesting.java. Simply:
    • edit the code, substituting your player for "RandomPlayer",
    • run the code (<= 11 hours), and
    • email me the last comma-separated-value (CSV) line of your output by June 15th, 2015 (Subject: “Parameterized Poker Squares Early Testing”).
    • Soon thereafter, I will distribute anonymous statistics on players’ performance for the different scoring systems of the testing.
• 11/19/14: How to introduce the Parameterized Poker Squares Challenge to your faculty and students:
  • PowerPoint and PDF presentations, Poker Squares and Contest handouts
  • Play the game together online or using a deck, pencil, and paper (grids only).
• Contact Todd Neller with team details if you're interested!
• 10/17/14: Contest description, draft code base for review (documentation)
• EAAI-14 Presentation (PPT)

Published results from contest:

• Karo Castro-Wunsch, William Maga, and Calin Anton, BeeMo, A Monte Carlo Simulation Agent for Playing Parameterized Poker Squares, in the Proceedings of the 30th AAAI Conference on Artificial Intelligence (AAAI-16), February 12–17, 2016, Phoenix, Arizona USA, AAAI Press, Palo Alto, California, USA, pp. 4071-4074. https://www.aaai.org/ocs/index.php/AAAI/AAAI16/paper/view/12405/12187 
• Todd W. Neller, Colin Messinger, and Zuozhi Yang, Learning and Using Hand Abstraction Values for Parameterized Poker Squares in the Proceedings of the 30th AAAI Conference on Artificial Intelligence (AAAI-16), February 12–17, 2016, Phoenix, Arizona USA, AAAI Press, Palo Alto, California, USA, pp. 4095-4100. http://cs.gettysburg.edu/~tneller/papers/eaai16.pdf
• Robert Arrington, Clay Langley, and Steven Bogaerts, Using Domain Knowledge to Improve Monte-Carlo Tree Search Performance in Parameterized Poker Squares, in the Proceedings of the 30th AAAI Conference on Artificial Intelligence (AAAI-16), February 12–17, 2016, Phoenix, Arizona USA, AAAI Press, Palo Alto, California, USA, pp. 4065-4070. https://www.aaai.org/ocs/index.php/AAAI/AAAI16/paper/view/11809/12186 
• Todd W. Neller, Zuozhi Yang, Colin M. Messinger, Calin Anton, Karo Castro-Wunsch, William Maga, Steven Bogaerts, Robert Arrington, and Clay Langley, Monte Carlo Approaches to Parameterized Poker Squares in Plaat, Aske, Kosters, Walter, van den Herik, Jaap (Eds.), LNCS: 9th International Conference, CG 2016, Leiden, The Netherlands, June 29 – July 1, 2016, Revised Selected Papers, Springer, Switzerland, 2016, pp. 22-33. http://cs.gettysburg.edu/~tneller/papers/cg2016.pdf

Talk slides:

• Todd W. Neller. Monte Carlo Approaches to Parameterized Poker Squares, The 9th International Conference on Computers and Games (CG2016), Leiden University, Leiden, the Netherlands, June 29th, 2016. http://cs.gettysburg.edu/~tneller/papers/talks/cg2016.pdf
• Todd W. Neller. The Parameterized Poker Squares EAAI NSG Challenge, The 30th AAAI Conference on Artificial Intelligence (AAAI-16), February 13, 2016. http://cs.gettysburg.edu/~tneller/papers/talks/eaai16PokerSquaresIntro.pdf
• Zuozhi Yang. Learning and Using Hand Abstraction Values for Parameterized Poker Squares, The 30th AAAI Conference on Artificial Intelligence (AAAI-16), February 13, 2016. http://cs.gettysburg.edu/~tneller/papers/talks/eaai16GettysburgPlayer.pdf

• Latest version of contest code (documentation)
• Main Poker Squares page
• Monte Carlo Tree Search (MCTS):
  • C. Browne et al. A Survey of Monte Carlo Tree Search Methods
  • L. Kocsis, C. Szepesvari. Bandit based Monte-Carlo Planning.
  • http://www.mcts.ai/?q=mcts
  • MCTS application to a similar problem:  R. Lorentz. An MCTS Program to Play EinStein Würfelt Nicht!

TinyURL for this page: http://tinyurl.com/ppokersqrs 

Todd W. Neller