Construction method of multidirectional neural network translation system

Abstract

The invention provides a construction method of a multidirectional neural network translation system. The construction method of the multi-directional neural network translation system is used for realizing machine translation operation based on a multi-directional neural network model. According to the construction method, sentence pair transformation processing about all source languages and alltarget languages is carried out on model training data; therefore, the model training data is updated, so that when the updated model training data is adopted to train the translation model, the translation model can perform direct decoding processing on the target language corresponding to any language translation direction, and a required translation result is obtained.

Description

Technical field [0001] The invention relates to the technical field of machine translation, in particular to a method for constructing a multi-directional neural network translation system. Background technique [0002] Machine translation is the use of a computer to convert a natural language as a source language into another natural language as a target language. Machine translation is widely used in text translation and speech translation. Users can translate between various natural languages ​​anytime and anywhere through terminals with machine translation functions. The existing machine translation methods are all based on neural network translation systems, but there is no unified neural network translation system that can simultaneously perform several different language translation operations. That is to say, a specific neural network translation system can only train the neural network translation model for a single language translation direction. This makes it necessar...

AI Technical Summary

Problems solved by technology

[0003] At present, machine translation programs require the user to first determine the language of the source language text and the language of the target language text before inputting the source language text, that is, the user needs to determine in advance the language translation direction corresponding to the current translation operation. If the user does not determine in advance There are many different language texts in the language translation direction or the source language text input by the user at the same time, which will cause the machine translation program to fail to perform translation operations normally to obtain the correct target language text

It can be seen that the existing machine translation method needs to deploy and train the neural network translation model for each language translation direction when there are many translation languages, which leads to the consumption of more development resources for this machine translation method; in addition, the current Some machine translation methods need to judge the language of the input source language text and determine the language translation direction corresponding to the current translation before running, and when the input source language text is mixed with other language texts at the same time, this machine translation method will not be able to obtain accurate translation. Translation results for

Images