Presentation is loading. Please wait.

Presentation is loading. Please wait.

Xi Breakthrough Player An extremely intelligent Breakthrough AI. You can tell by the use of X. CSE 486 B Fall 2012 Miami University.

Similar presentations


Presentation on theme: "Xi Breakthrough Player An extremely intelligent Breakthrough AI. You can tell by the use of X. CSE 486 B Fall 2012 Miami University."— Presentation transcript:

1 Xi Breakthrough Player An extremely intelligent Breakthrough AI. You can tell by the use of X. CSE 486 B Fall 2012 Miami University

2 Team Mike Jacobs Jiang Nuo Reuben Smith

3 Components Search Threading Depth control Storage optimization Features Parametric feature weights Feature weight optimization Opponent interfacing

4 Search Negamax with alpha-beta – Negamax is minimax “simplified” – State evaluation is from the perspective of the player associated with the current depth – Scores are negated as search moves up the tree

5 Threading Threading is used to split up the search workload evenly between search workers – The number of workers created is based on the number of processor cores available less one Workers use copies of state, so no concurrency issues are possible

6 Depth Control We dynamically adjust depth limit after each move targetMoveTime = gameTimeLimit / averageGameMoves – targetMoveTime is calculated at the start of each game – averageGameMoves is updated after each game

7 Depth Control, cont. If search finishes in less than targetMoveTime, then the depth limit increases If search finishes in more than targetMoveTime, then the depth limit decreases Depth control uses coefficients to affect how easily the depth limit can be changed again This allows search depth to stabilize based on system performance and search domain size

8 Storage Optimization Instead of creating a move object for each possible move at every node in the search tree, move data is encoded inside a primitive short: – Bits 0-2 represent starting row – Bits 3-5 represent starting column – Bits 6-8 represent ending row – Bits 9-11 represent ending column – Bits 12-15 represent move flags (e.g. move is a take)

9 Features 7 features implemented Notable features: – Squares owned – Penetration – Spread – Conflict – Cover

10 Parametric Feature Weights The player can take in a parameter object to allow user-defined weights This can be used to control the difficulty level of the player; for example: – easy – only one feature is “on” (has weight) – hard – all features are given an optimal weighting

11 Feature Weight Optimization Hill climbing can be used in two different places to optimize feature weights: 1.To determine optimal weight for each feature in a set of features 2.To choose the set of features that are turned on (from sets of 2 to 4 features) – Results from #1 would be used to find a maximum for scoring in #2 – Momentum would be used to improve the chances of finding the global maximum This was not included in the final submission due to time and computational requirements

12 Opponent Interfacing (taunting) Lastly, the player will also maintain superiority by intimidating the opponent at the start of each game: – “Hello,. Nice to beat you.“ – “Spill the blood of the innocent!”

13 ?


Download ppt "Xi Breakthrough Player An extremely intelligent Breakthrough AI. You can tell by the use of X. CSE 486 B Fall 2012 Miami University."

Similar presentations


Ads by Google