Planning to Guide Opening and Middle Game Play in Shogi Reijer Grimbergen (Electrotechnical Laboratory) Hitoshi Matsubara (Future University Hakodate)

Slides:



Advertisements
Similar presentations
Adversarial Search Chapter 6 Sections 1 – 4. Outline Optimal decisions α-β pruning Imperfect, real-time decisions.
Advertisements

METAGAMER: An Agent for Learning and Planning in General Games Barney Pell NASA Ames Research Center.
Adversarial Search Chapter 6 Section 1 – 4. Types of Games.
February 7, 2006AI: Chapter 6: Adversarial Search1 Artificial Intelligence Chapter 6: Adversarial Search Michael Scherger Department of Computer Science.
September 8th 2005Advances in Computer Games 111 Enhancing Search Efficiency by Using Move Categorization Based on Game Progress in Amazons Yoshinori Higashiuchi.
ICS-271:Notes 6: 1 Notes 6: Game-Playing ICS 271 Fall 2008.
CS 484 – Artificial Intelligence
2002/11/15Game Programming Workshop1 A Neural Network for Evaluating King Danger in Shogi Reijer Grimbergen Department of Information Science Saga University.
Adversarial Search Chapter 5.
COMP-4640: Intelligent & Interactive Systems Game Playing A game can be formally defined as a search problem with: -An initial state -a set of operators.
Application of Artificial intelligence to Chess Playing Capstone Design Project 2004 Jason Cook Bitboards  Bitboards are 64 bit unsigned integers, with.
Adversarial Search CSE 473 University of Washington.
Artificial Intelligence for Games Game playing Patrick Olivier
An Introduction to Artificial Intelligence Lecture VI: Adversarial Search (Games) Ramin Halavati In which we examine problems.
1 Adversarial Search Chapter 6 Section 1 – 4 The Master vs Machine: A Video.
Games CPSC 386 Artificial Intelligence Ellen Walker Hiram College.
G51IAI Introduction to AI Minmax and Alpha Beta Pruning Garry Kasparov and Deep Blue. © 1997, GM Gabriel Schwartzman's Chess Camera, courtesy IBM.
Selective Search in Games of Different Complexity Maarten Schadd.
Minimax and Alpha-Beta Reduction Borrows from Spring 2006 CS 440 Lecture Slides.
October 28th 2001GPW20011 Using Castle and Assault Maps for Guiding Opening and Middle Game Play in Shogi Reijer Grimbergen (Saga University) Jeff Rollason.
Plausible Move Generation Using Move Merit Analysis in Shogi Reijer Grimbergen (Electrotechnical Laboratory) Hitoshi Matsubara (Future University Hakodate)
UNIVERSITY OF SOUTH CAROLINA Department of Computer Science and Engineering CSCE 580 Artificial Intelligence Ch.6: Adversarial Search Fall 2008 Marco Valtorta.
ICS-271:Notes 6: 1 Notes 6: Game-Playing ICS 271 Fall 2006.
Othello Sean Farrell June 29, Othello Two-player game played on 8x8 board All pieces have one white side and one black side Initial board setup.
Game Playing State-of-the-Art  Checkers: Chinook ended 40-year-reign of human world champion Marion Tinsley in Used an endgame database defining.
CSC 412: AI Adversarial Search
Game Playing.
Introduction Many decision making problems in real life
Othello Artificial Intelligence With Machine Learning
Agents that can play multi-player games. Recall: Single-player, fully-observable, deterministic game agents An agent that plays Peg Solitaire involves.
2004/11/13GPW20041 What Shogi Programs Still Cannot Do - A New Test Set for Shogi - Reijer Grimbergen and Taro Muraoka Department of Informatics Yamagata.
Adversarial Search Chapter 6 Section 1 – 4. Outline Optimal decisions α-β pruning Imperfect, real-time decisions.
Introduction to Artificial Intelligence CS 438 Spring 2008 Today –AIMA, Ch. 6 –Adversarial Search Thursday –AIMA, Ch. 6 –More Adversarial Search The “Luke.
2009/11/14GPW20091 Analysis of the Behavior of People Solving Sudoku Puzzles Reijer Grimbergen School of Computer Science, Tokyo University of Technology.
Othello Playing AI Matt Smith. Othello 8x8 Board game 8x8 Board game Try to outflank opponents pieces Try to outflank opponents pieces Winner ends up.
1 Phase II - Checkers Operator: Eric Bengfort Temporal Status: End of Week Five Location: Phase Two Presentation Systems Check: Checkers Checksum Passed.
Application of Artificial Intelligence to Chess Playing Jason Cook Capstone Project.
Memory and Analogy in Game-Playing Agents Jonathan Rubin & Ian Watson University of Auckland Game AI Group
Games. Adversaries Consider the process of reasoning when an adversary is trying to defeat our efforts In game playing situations one searches down the.
1 Adversarial Search CS 171/271 (Chapter 6) Some text and images in these slides were drawn from Russel & Norvig’s published material.
Draughts. Introduction Draughts is played on the same chequered board as chess and has been played in Europe since the end of the 1100s. It is similar.
CHAPTER 4 PROBABILITY THEORY SEARCH FOR GAMES. Representing Knowledge.
INTELLIGENT SYSTEM FOR PLAYING TAROK
IWEC20021 Threat Stacks to Guide Pruning and Search Extensions in Shogi Reijer Grimbergen Department of Information Science Saga University, Japan.
PYIWIT'021 Threat Analysis to Reduce the Effects of the Horizon Problem in Shogi Reijer Grimbergen Department of Information Science Saga University.
GAME PLAYING 1. There were two reasons that games appeared to be a good domain in which to explore machine intelligence: 1.They provide a structured task.
Adversarial Search Chapter Games vs. search problems "Unpredictable" opponent  specifying a move for every possible opponent reply Time limits.
Each piece is represented by a symbol. The pieces all stand in the same position at the start of the game the pieces are the Rook, the Knight, the Bishop,
2008/09/30Computers and Games Cognitive Modeling of Knowledge-Guided Information Acquisition in Games Reijer Grimbergen Department of Informatics.
Computers and Games Board Maps and Hill-Climbing for Opening and Middle Game Play in Shogi Reijer Grimbergen (Saga University) Jeff Rollason (Oxford.
2008/11/08GPW20081 A Reproduction Experiment Concerning the Relation Between Perceptual Features and Memory in Shogi Reijer Grimbergen Department of Informatics.
Othello Artificial Intelligence With Machine Learning Computer Systems TJHSST Nick Sidawy.
Let’s----- go play now. Let’s--- go--- play now. Let’s go play--- now---. Let’s go--- play now.
Using Castle Maps for Guiding Opening and Middle Game Play in Shogi Reijer Grimbergen (Saga University) Jeff Rollason (Oxford Softworks)
Explorations in Artificial Intelligence Prof. Carla P. Gomes Module 5 Adversarial Search (Thanks Meinolf Sellman!)
Plausible Move Generation Using Move Merit Analysis with Cut-off Thresholds in Shogi Reijer Grimbergen (Electrotechnical Laboratory)
Adversarial Search Chapter 5 Sections 1 – 4. AI & Expert Systems© Dr. Khalid Kaabneh, AAU Outline Optimal decisions α-β pruning Imperfect, real-time decisions.
ADVERSARIAL SEARCH Chapter 6 Section 1 – 4. OUTLINE Optimal decisions α-β pruning Imperfect, real-time decisions.
Adversarial Search CMPT 463. When: Tuesday, April 5 3:30PM Where: RLC 105 Team based: one, two or three people per team Languages: Python, C++ and Java.
Game Playing Why do AI researchers study game playing?
Adversarial Search Chapter 5.
Using Bitboards for Move Generation in Shogi
Adversarial Search.
On the Relation Between Perception, Memory and Cognition in Games
Counting to 100 Counting by ones
Adversarial Search CMPT 420 / CMPG 720.
Adversarial Search CS 171/271 (Chapter 6)
Othello Artificial Intelligence With Machine Learning
Adversarial Search Chapter 6 Section 1 – 4.
Presentation transcript:

Planning to Guide Opening and Middle Game Play in Shogi Reijer Grimbergen (Electrotechnical Laboratory) Hitoshi Matsubara (Future University Hakodate)

Opening Databases in Two-player Complete Information Games Two-player complete information games: Starting position determined by the rules of the game. Opening book: Database from which early moves can be played. Vital in chess, checkers and Othello programs. Our opening book for Shogi: 1000 professional games. More than 20 books on joseki. More than 110,000 positions

Opening Book Use in Shogi We tested the effectiveness of our opening book in 25 games against AI Shogi 2000, Kakinoki Shogi IV, Todai Shogi 2 and Kanazawa Shogi 98:

Opening Book Use in Shogi Our program gets out of book quickly: Within five moves in 32 games; within ten moves in 71 games. Average: out of book after 8.5 moves. Most of the knowledge in the opening book is never used. Idea: use opening book also when there is no perfect match between the position on the board and the position in the opening book

Partial Matching Half matching: all of the pieces of the player to move in the current position are on the same squares as the pieces in a position in the database. Quarter matching: look only at pieces in two areas of the shogi board: i

Implementation If there is a full match, book move is played. No full match: both half and quarter matching. Half matching: current move number equals the move number of the position in the opening book. Quarter matching: move difference between current position and opening book position is smaller than 10. Bonuses based on the number of times the move was found: Half matching: max(0.75*PV, (0.5*PV+2*FQ)) Quarter matching: max(0.25*PV, (0.125*PV+FQ))

Experiments Opening book extension with partial matching: how far can the use of the opening book be extended by using our method? Self-play experiments between three identical shogi programs: No partial matching. Half matching. Half matching and quarter matching.

Extension Experiment Same 25 games against AI Shogi 2000, Kakinoki Shogi IV, Kanazawa Shogi 98 and Todai Shogi 2: OpponentFullHalfQuarter #Mv Played#MvPlayed AI Shogi % % Kakinoki Shogi IV % % Kanazawa Shogi % % Todai Shogi % % Total % %

Self-Play Experiment 100 games between three identical shogi programs: NM: No Matching. HM: Half Matching. HQM: Half and Quarter Matching. NoVersion123PWL 1HQMX HM46-54X HN X073127

Related Work Match only with pieces in one’s own camp (Kotani 1999). Generate move sequences from an opening book (Nakaie 1996). Pattern matching with opening book (Nakaie 1997).

Conclusions and Future Work Conclusions: Half matching and quarter matching significantly extends the phase of the game in which the opening book can be used. Half matching and quarter matching improves the playing strength of a shogi program. Future work: Tuning of the bonus values. Finding the optimal quarter match window. Investigate other types of matching: e.g. also match the pieces of the opponent, match even smaller areas, use of similarity count.