A Data-Driven Model for Automated Chinese Word Segmentation and POS Tagging

Table 3

Parameters of the improved AlexNet.

Layer	Convolution kernel size	Step length	Number of cores
Convolutional layer 1	11 × 11 × 3	4	96
Pooling layer 1	1 × 3 × 3 × 1	2	—
Convolutional layer 2	5 × 5 × 48	1	256
Pooling layer 2	1 × 3 × 3 × 1	2	—
Convolution layer 3	3 × 3 × 384	1	256
Convolutional layer 4	3 × 64 × 384	1	384
Convolutional layer 5	3 × 3 × 256	1	384
Pooling layer 5	1 × 3 × 3 × 1	2	—
Fully connected layer 6	32 × 256	—	256
ReLU6	—	—	384
Dropout layer	—	—	—
Fully connected layer 7	32 × 256	—	256
ReLU7	—	—	256
Dropout layer	—	—	—
Fully connected layer 8	256 × 2	—	2
Prob	Softmax	—	—
Output	—	—	—