Machine translation automatic post-editing method and device

Abstract

The embodiment of the invention provides a machine translation automatic post-editing method and device. The method comprises the steps that a target source language sentence and a target machine translation draft are obtained; the target source language sentence and the target machine translation draft are input into an automatic post-editing deep learning model, and an edited sentence corresponding to the target machine translation draft is output; wherein the automatic post-editing deep learning model is generated after training by using a deep neural network for copy explicit modeling based on a training set containing copy tags; wherein the training set containing the copy label comprises a preset number of training samples; wherein each training sample is a quadruple comprising a sample source language sentence, a sample machine translation draft, a sample edited sentence and a copy label sequence, and the copy label sequence is formed by arranging copy labels of all words in thesample machine translation draft according to the sequence of the words in the corresponding sample machine translation draft. The editing effect of automatic post-editing of machine translation canbe improved, and the accuracy is high.

Description

technical field [0001] The invention relates to the technical field of machine translation, in particular to a method and device for automatic post-editing of machine translation. Background technique [0002] With the deepening of international exchanges, people's demand for language translation is increasing day by day. However, there are many kinds of languages ​​in the world, each with its own characteristics and flexible forms, making the automatic processing of languages, including machine translation between languages, a crucial technology. [0003] Machine translation is the translation of text from one specific language to another by a computer without human guidance. However, the quality of translation is not satisfactory. In order to ensure the quality of translation, it is necessary to manually edit and modify the translated text (that is, the machine translation draft). Such post-translation editing consumes less manpower than directly translating the source la...

AI Technical Summary

Problems solved by technology

[0005] However, the existing automatic post-editing uses sequence-to-sequence methods based on deep learning, although post-translation editing can be done, but the accuracy is low and the editing effect is poor

Images