COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview...
Transcript of COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview...
![Page 1: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/1.jpg)
COMP219:Artificial Intelligence
Lecture 13: Game Playing
1
![Page 2: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/2.jpg)
Overview
• Last time– Search with partial/no observations– Belief states– Incremental belief state search– Determinism vs non-determinism
• Today– We will look at how search can be applied to playing games
• Types of games• Perfect play
– minimax decisions– alpha–beta pruning
• Playing with limited recourses
2
![Page 3: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/3.jpg)
Games and Search
• In search we make all the moves. In games we play against an “unpredictable” opponent– Solution is a strategy specifying a move for every possible
opponent reply• Assume that the opponent is intelligent: always makes
the best move• Some method is needed for selecting good moves that
stand a good chance of achieving a winning position, whatever the opponent does!
• There are time limits, so we are unlikely to find goal, and must approximate using heuristics
3
![Page 4: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/4.jpg)
Types of Game
• In some games we have perfect information – the position is known completely
• In others we have imperfect information: e.g. we cannot see the opponent’s cards
• Some games are deterministic – no random element
• Others have elements of chance (dice, cards)
4
![Page 5: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/5.jpg)
Types of Games
5
![Page 6: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/6.jpg)
We will consider:
• Games that are:– Deterministic– Two-player– Zero-sum
• the utility values at the end are equal and opposite• example: one wins (+1) the other loses (−1)
– Perfect information• E.g. Othello, Blitz Chess
6
![Page 7: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/7.jpg)
Problem Formulation
• Initial state– Initial board position, player to move
• Transition model– List of (move, state) pairs, one per legal move
• Terminal test– Determines when the game is over
• Utility function– Numeric value for terminal states– e.g. Chess +1, -1, 0– e.g. Backgammon +192 to -192
7
![Page 8: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/8.jpg)
Noughts and Crosses
X X
O X
O O
8
![Page 9: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/9.jpg)
Noughts and CrossesX X
O X
O O X
X X
O X
O O
X X X
O X
O O
X X
O X X
O O
9
![Page 10: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/10.jpg)
Noughts and CrossesX X
O X
O O X
X X
O X
O O
X X X
O X
O O
X X
O X X
O O
+1
+1
10
![Page 11: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/11.jpg)
Noughts and CrossesX X
O X
O O X
X X
O X
O O
X X X
O X
O O
X X
O X X
O O
X X
O X X
O O O
X O X
O X X
O O
+1
+1
11
![Page 12: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/12.jpg)
Noughts and CrossesX X
O X
O O X
X X
O X
O O
X X X
O X
O O
X X
O X X
O O
X X
O X X
O O O
X O X
O X X
O O
+1
+1
-1
12
![Page 13: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/13.jpg)
Noughts and CrossesX X
O X
O O X
X X
O X
O O
X X X
O X
O O
X X
O X X
O O
X X
O X X
O O O X O X
O X X
O O X
X O X
O X X
O O
+1
+1
-1
13
![Page 14: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/14.jpg)
Noughts and CrossesX X
O X
O O X
X X
O X
O O
X X X
O X
O O
X X
O X X
O O
X X
O X X
O O O X O X
O X X
O O X
X O X
O X X
O O
+1
+1
-1
+1
14
![Page 15: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/15.jpg)
Game Tree
• Each level labelled with player to move• Each level represents a ply
– Half a turn• Represents what happens with competing
agents
15
![Page 16: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/16.jpg)
Introducing MIN and MAX
• MIN and MAX are two players:– MAX wants to win (maximise utility)– MIN wants MAX to lose (minimise utility for MAX)– MIN is the Opponent
• Both players will play to the best of their ability– MAX wants a strategy for maximising utility assuming
MIN will do best to minimise MAX’s utility– Consider minimax value of each node
16
![Page 17: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/17.jpg)
Example Game Tree
Minimax value of a nodeis the value of the best terminal node, assumingBest play by opponent
17
![Page 18: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/18.jpg)
Minimax Value
18
• Utility for MAX of being in that state assuming both players play optimally to the end of the game
• Formally:
![Page 19: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/19.jpg)
Minimax Algorithm
• Calculate minimaxValue of each node recursively
• Depth-first exploration of tree• Game tree as minimax tree• Max Node:
• Min Node
19
![Page 20: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/20.jpg)
Minimax Tree
20
![Page 21: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/21.jpg)
Minimax Tree
3 12 8
21
![Page 22: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/22.jpg)
Minimax Tree
3 12 8
3
Min takes the lowest value from its children
22
![Page 23: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/23.jpg)
Minimax Tree
3 12 8
3
2 4 6
Min takes the lowest value from its children
23
![Page 24: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/24.jpg)
Minimax Tree
3 12 8
3
2 4 6
2
Min takes the lowest value from its children
24
![Page 25: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/25.jpg)
Minimax Tree
3 12 8
3
2 4 6
2
14 5 2
Min takes the lowest value from its children
25
![Page 26: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/26.jpg)
Minimax Tree
3 12 8
3
2 4 6
2
14 5 2
2
Min takes the lowest value from its children
26
![Page 27: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/27.jpg)
Minimax Tree
3 12 8
3
2 4 6
2
14 5 2
2
3
Min takes the lowest value from its children
27
Max takes the highest value from its children
![Page 28: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/28.jpg)
Exercise
28
• Perform the minimax search algorithm on the following tree to get the minimax value of the root:
![Page 29: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/29.jpg)
Exercise
29
6 4 7 6 8 11 1 2
![Page 30: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/30.jpg)
Exercise
30
6 4 7 6 8 11 1 2
6 7 11 2
26
6
![Page 31: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/31.jpg)
Properties of Minimax
• Complete, if tree is finite• Optimal, against an optimal opponent. Otherwise??
– No. e.g. expected utility against random player• Time complexity: bm
• Space complexity: bm (depth-first exploration)– For chess, b ≈ 35, m ≈ 100 for “reasonable” games– Infeasible – so typically set a limit on look ahead. Can still use
minimax, but the terminal node is deeper on every move, so there can be surprises. No longer optimal
• But do we need to explore every path?
31
![Page 32: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/32.jpg)
Pruning
• Basic idea:If you know half-way through a calculation that it will succeed or fail, then there is no point in doing the rest of it
• For example, in Java it is clear that when evaluating statements like
if ((A > 4) || (B < 0))• If A is 5 we do not have to check on B!
32
![Page 33: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/33.jpg)
Alpha-Beta Pruning
33
![Page 34: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/34.jpg)
Alpha-Beta Pruning
3 12 8
34
![Page 35: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/35.jpg)
Alpha-Beta Pruning
3 12 8
3
35
![Page 36: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/36.jpg)
Alpha-Beta Pruning
3 12 8
3
≥ 3
36
![Page 37: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/37.jpg)
Alpha-Beta Pruning
3 12 8
3
2
≥ 3
37
![Page 38: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/38.jpg)
Alpha-Beta Pruning
3 12 8
3
2
≤ 2
≥ 3
38
X X
![Page 39: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/39.jpg)
Alpha-Beta Pruning
3 12 8
3
2
≤ 2
14
≥ 3
39
X X
![Page 40: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/40.jpg)
Alpha-Beta Pruning
3 12 8
3
2
≤ 2
14
≥ 3
40
≤ 14
X X
![Page 41: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/41.jpg)
Alpha-Beta Pruning
3 12 8
3
2
≤ 2
14 5
≥ 3
41
≤ 14
X X
![Page 42: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/42.jpg)
Alpha-Beta Pruning
3 12 8
3
2
≤ 2
14 5
≥ 3
42
≤ 5
X X
![Page 43: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/43.jpg)
Alpha-Beta Pruning
3 12 8
3
2
≤ 2
14 5 2
≥ 3
43
≤ 5
X X
![Page 44: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/44.jpg)
Alpha-Beta Pruning
3 12 8
3
2
≤ 2
14 5 2
≥ 3
44
2
X X
![Page 45: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/45.jpg)
Alpha-Beta Pruning
3 12 8
3
2
≤ 2
14 5 2
≥ 3
45
2
X X
![Page 46: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/46.jpg)
Why is it called alpha-beta?
46
![Page 47: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/47.jpg)
The Alpha-Beta Algorithm
• alpha (α) is value of best (highest value) choice for MAX
• beta (β) is value of best (lowest value) choice for MIN
• If at a MIN node and value ≤ α, stop looking, because MAX node will ignore this choice
• If at a MAX node and value ≥ β beta, stop looking because MIN node will ignore this choice
47
![Page 48: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/48.jpg)
Properties of Alpha-Beta
• Pruning does not affect final result• Good move ordering improves effectiveness of
pruning• With “perfect ordering” time complexity bm/2 and
so doubles solvable depth• A simple example of the value of reasoning about
which computations are relevant (a form of meta-reasoning)
• Unfortunately, 3550 is still impossible, so chess not completely soluble
48
![Page 49: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/49.jpg)
Cutoffs and Heuristics• Cut off search according to some cutoff test
– Simplest is a depth limit• Problem: payoffs are defined only at terminal states• Solution: Evaluate the pre-terminal leaf states using heuristic evaluation
function rather than using the actual payoff function
49
![Page 50: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/50.jpg)
Cutoff Value
• To handle the cutoff, in minimax or alpha-beta search we can make an alteration by making use of a cutoff value
• MinimaxCutof is identical to MinimaxValue
except1. Terminal test is replaced by Cutof test, which indicates
when to apply the evaluation function2. Utility is replaced by Evaluation function, which
estimates the position’s utility
50
![Page 51: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/51.jpg)
Example: Chess (I)• Assume MAX is white• Assume each piece has the following material value:
– pawn = 1– knight = 3– bishop = 3– rook = 5– queen = 9
• let w = sum of the value of white pieces• let b = sum of the value of black pieces
51
![Page 52: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/52.jpg)
Example: Chess (II)
• The previous evaluation function naively gave the same weight to a piece regardless of its position on the board...– Let Xi be the number of squares the i-th piece attacks– Evaluation(n) = piece1value * X1 + piece2value * X2 + ...
52
![Page 53: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/53.jpg)
Example: Chess (III)
• Heuristics based on database search– Statistics of wins in the position under
consideration– Database defining perfect play for all positions
involving X or fewer pieces on the board (endgames)
– Openings are extensively analysed, so can play the first few moves “from the book”
53
![Page 54: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/54.jpg)
Deterministic Games in Practice
• Draughts: Chinook ended 40-year-reign of human world champion Marion Tinsley in 1994. Used an endgame database defining perfect play for all positions involving 8 or fewer pieces on the board, a total of 443,748,401,247 positions
• Chess: Deep Blue defeated human world champion Gary Kasparov in a six-game match in 1997. Deep Blue searches 200 million positions per second, used very sophisticated evaluation, and undisclosed methods for extending some lines of search up to 40 ply
54
![Page 55: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/55.jpg)
Deterministic Games in Practice
• Othello: human champions refuse to compete against computers, who are too good
• Go: a challenging game for AI (b > 300) so progress much slower with computers. AlphaGo was a recent breakthrough
55
See more at: University of Alberta GAMES Group
![Page 56: COMP219: Artificial Intelligence · Artificial Intelligence Lecture 13: Game Playing 1. Overview • Last time – Search with partial/no observations – Belief states – Incremental](https://reader034.fdocuments.in/reader034/viewer/2022042402/5f11cc0b3785533f8752a191/html5/thumbnails/56.jpg)
Summary
• Games have been an AI topic since the beginning. They illustrate several important features of AI:– perfection is unattainable so must approximate– good idea to think about what to think about– uncertainty constrains the assignment of values to states – optimal decisions depend on information state, not real state
• Next lecture:– We have now finished with the topic of search so we will
move on to knowledge representation
56