Part-of-Speech Tagging of English-Translation Bilingual Parallel Corpus Based on Convolutional Neural Network
Xiuli Gou, Rui Li, Peng Zhang · 2022 IEEE Conference on Telecommunications, Optics and Computer Science (TOCS) · 2022
Annotation is the process of adding information to a corpus. The key to realizing the machine-reading of corpus and improving the value of corpus utilization is effective annotation. The convolutional neural network and its mathematical model are constructed, and the network structure, convolutional layer and pooling layer, activation function, loss function and optimization are studied in detail; a bilingual parallel corpus part-of-speech tagging model is constructed, including defining the objective function, determining the parameters and parameter space; the training process of the convolutional neural network is designed, including the forward propagation stage and the back propagation stage. Based on the convolutional neural network annotation, the corresponding features are learned from a large number of samples, which reduces the complexity of the network model, reduces the number of weights, avoids the complex feature extraction process, and promotes the research and application of machine translation.