A Study on Chinese-English Machine Translation Based on Transfer Learning and Neural Networks

Table 3

Experimental results of the recurrent neural network-based monolingual to machine translation quality assessment model under nonjoint training conditions.

Method	Data	Pearson’s	MAE	RMSE
Transformer+embedding+no-fine-tuning	Dev	0.4628	0.1385	0.1257
	Test2016	0.4346	0.1222	0.1529
	Test2017	0.4046	0.1082	0.1659
Transformer+latent+ no-fine-tuning	Dev	0.6314	0.1067	0.1513
	Test2016	0.8471	0.1017	0.1125
	Test2017	0.6401	0.1321	0.1158
Transformer+wordpredicting+ no-fine-tuning	Dev	0.6965	0.0789	0.1876
	Test2016	0.764	0.0794	0.1268
	Test2017	0.6956	0.1571	0.1359