Mapping Relationship Discovery of Multidimensional Architectures in Autonomous Transportation System Based on Text-Matching Model

Table 5

Parameters of text-matching model.

Text-matching model	CNN		Number of LSTM units	Dropout rate	Optimizer	Number of epochs	Batch size
	Number of filters	Size of kernels
ABCNN	50	3	—	0.3	Adam (1e − 3)	50	200
Siamese-Bi-LSTM	—	—	F: 100	0.2	Adam (5e − 4)	100	256
			B: 100
ESIM	—	—	F: 256	0.3	Adam (5e − 4)	50	50
			B: 256
BERT-wwm-CLS	—	—	—	0.1	Adam (2e − 5)	20	32
BERT-wwm-Bi-LSTM	—	—	F: 100	0.1	Adam (2e − 5)	20	32
			B: 100
BERT-wwm-CNN	128	3,4,5	—	0.1	Adam (2e − 5)	20	32
Siamese-BERT-wwm-CLS	—	—	—	0.1	Adam (2e − 5)	20	32
Siamese-BERT-wwm-CNN	128	3,4,5	—	0.1	Adam (2e − 5)	20	32
Siamese-BERT-wwm-Bi-LSTM-CNN	128	3,4,5	F: 100	0.1	Adam (2e − 5)	20	32
			B: 100
Siamese-BERT-wwm-CNN-Bi-LSTM	128	3,4,5	F: 100	0.1	Adam (2e − 5)	20	32
			B: 100
Siamese-BERT-wwm-CNN + Bi-LSTM	128	3,4,5	F: 100	0.1	Adam (2e − 5)	20	32
			B: 100
Siamese-BERT-Bi-LSTM	—	—	F: 100	0.1	Adam (2e − 5)	20	32
			B: 100
Siamese-ALBERT-Bi-LSTM	—	—	F: 100	0.1	Adam (2e − 5)	20	32
			B: 100
Siamese-roberta-Bi-LSTM	—	—	F: 100	0.1	Adam (2e − 5)	20	32
			B: 100
Siamese-BERT-wwm-Bi-LSTM	—	—	F: 100	0.1	Adam (2e − 5)	20	32
			B: 100