0.25
0.5
0.75
1.25
1.5
1.75
2
Trade-Offs in Sampling-Based Adversarial Planning
Published on Jul 21, 20113692 Views
The Upper Confidence bounds for Trees (UCT) algorithm has in recent years captured the attention of the planning and game-playing community due to its notable success in the game of Go. However, attem
Related categories
Chapter list
Trade-offs in Sampling-based Adversarial Planning00:00
Upper Confidence bounds for Trees (UCT)00:11
Understanding UCT00:42
The Multi-Armed Bandit Problem - 201:26
The UCB1 Bandit Algorithm - 101:31
The UCB1 Bandit Algorithm - 202:59
From Bandits to Tree Search03:45
The UCT Algorithm - 104:21
The UCT Algorithm - 205:47
UCT in Action06:26
Minimax in Action07:17
UCT versus Minimax - 107:42
UCT versus Minimax - 208:30
Mancala09:04
UCT in Mancala09:51
Complete versus Selective Search - 111:56
Complete versus Selective Search - 212:23
Other Trade-offs in UCT14:08
UCTMAXh versus Minimax15:40
Background: Trap States - 116:13
Background: Trap States - 216:52
Traps in Mancala17:07
‘Partial’ Games of Mancala17:39
UCTMAXh versus Minimax - 318:05
Conclusions19:24