Improved Feature Learning: A Maximum-Average-Out Deep Neural Network for the Game Go

Table 2

Test results.

Comparison	Black side	White side	Result
1	MAO	ResNet18	B + resign
2	ResNet18	MAO	W + resign
3	MAO	ResNet18	B + resign
4	ResNet18	MAO	W + resign
5	MAO	ResNet18	W + resign
6	ResNet18	MAO	W + resign
7	MAO	ResNet18	B + resign
8	ResNet18	MAO	W + resign
9	MAO	ResNet18	B + resign
10	ResNet18	MAO	B + resign