English-Burmese bilingual parallel sentence pair extraction method and device based on BiLSTM-CNN

Abstract

The invention relates to an English-Burmese bilingual parallel sentence pair extraction method and device based on BiLSTM-CNN, and belongs to the technical field of natural language processing. The method comprises the following steps: firstly, pre-training a bilingual word vector through a Muse tool; secondly, performing function marking on the sentence by utilizing the characteristics of the Burmese virtual words and the Burmese assistant words for identifying the subject-called guest of the Burmese, splicing syntactic structure information of each word into a word vector, encoding the sentence by using BiLSTM-CNN, and taking an output probability as a condition for measuring whether the sentence is a parallel sentence pair or not. According to the above steps, the BiLSTM-CNN-based British-Burmese bilingual parallel sentence pair extraction device is prepared through functional modularization. Compared with a traditional bilingual parallel sentence pair recognition system, the methodand the device are simpler. Experimental results show that the method and the device are superior to a baseline system in the aspects of accuracy, recall rate and other indexes. The accuracy is generally improved.

Description

technical field [0001] The invention relates to a BiLSTM-CNN-based English-Myanmar bilingual parallel sentence pair extraction method and device, and belongs to the technical field of natural language processing. Background technique [0002] In the field of natural language processing, the scale of parallel corpora plays an important role in improving the performance of machine translation. For Burmese, which is resource-scarce, the English-Myanmar parallel corpus resources are seriously scarce, and the quality of machine translation has not yet reached a practical level. The traditional methods of obtaining parallel corpora include manual translation and machine translation. The former is more costly and inefficient, while the latter relies on machine translation performance and poor quality. The scale of parallel corpus on the Internet is relatively small, while the comparable corpus is relatively large. How to use the large amount of English-Myanmar comparable corpus on ...

AI Technical Summary

Problems solved by technology

Traditional methods often only consider the semantic information of their own language. In fact, different languages ​​contain corresponding functional structures. Since the semantic expression of sentences is closely related to the syntactic structure, although the existing representation methods can retain the Word order information, but the loss of syntactic structure information cannot be avoided, and it is difficult to accurately learn sentence representation

Images