Method and system for extracting resequencing template in machine translation

Abstract

The invention relates to a method and a system for extracting a resequencing template in machine translation. The method comprises the following steps of: (1) inputting bilingual alignment corpus and carrying out word segmentation and part-of-speech tagging on a source language part in the bilingual alignment corpus; (2) resequencing and analyzing each bilingual pair of sentences in the bilingual alignment corpus and extracting a resequencing example; (3) dividing each resequencing example into two parts according to the positions of a word pair in a source language and a target language in the resequencing example, and for each part, confirming a variable part according to the part-of-speech tagging, and replacing the variable part with a variable. The invention can eliminate the restriction to extract a translation template in the prior art and can extract various resequencing templates so as to increase the coverage rate of the resequencing template on resequencing phenomena in translation.

Description

technical field [0001] The invention relates to the field of machine translation, in particular to a method and system for extracting a sequence template in machine translation. Background technique [0002] A translation template is a knowledge representation commonly used in machine translation to guide translation, and describes the corresponding relationship that needs to be followed when translating from a source language to a target language. The translation template is a string composed of constants and variables on the source language side and the target language side, and each part of the source language and target language strings corresponds one-to-one. [0003] An example of a simple Chinese-English translation template: [0004] X today. [0005] X today. [0006] The constants in the template refer to language fragments, also known as terminators. For example, "today" in the above example corresponds to "today"; "." corresponds to ".". A variable refers to ...

AI Technical Summary

Problems solved by technology

However, the translation templates used in the machine translation system in the prior art cannot accurately and completely describe the reordering phenomenon due to various limitations in the extraction process; When the translation template is used, it will be translated sequentially by default, resulting in wrong ordering

[0011] Therefore, the existing automatic extraction methods for sequencing templates have the problem of low coverage of sequencing templates for sequencing phenomena in translation.

Images