Seminar: Monte Carlo Tree Search for the Imperfect Information Card Game Cribbage
Supervisor: Dr. David Churchill
Monte Carlo Tree Search for the Imperfect Information Card Game Cribbage
Department of Computer Science
Tuesday, April 4, 2017, 12:00 p.m., Room EN-2022
Monte Carlo Tree Search (MCTS), an AI technique that uses random sampling of game playouts, has been used successfully in perfect information deterministic games such as Go since 2006. It is a useful game AI technique because it doesn't rely on evaluations of non-terminal game states or other domain knowledge.
Several techniques for handling imperfect information and stochastic game elements in MCTS have been used. In this research, we implemented a Cribbage card game for two players and investigated several methods for handling imperfect information for an MCTS AI player.