Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Machine translation method, training method, corresponding device and electronic equipment

A machine translation and target language technology, applied in the field of natural language processing, can solve problems such as consumption of computing resources, high labor costs, and no improvement in algorithms, so as to improve processing efficiency, avoid multiple calculations, and reduce the amount of calculation.

Inactive Publication Date: 2020-04-03
BEIJING SAMSUNG TELECOM R&D CENT +1
View PDF0 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the method of parallel computing does not improve the algorithm itself, and it needs to consume more computing resources; program optimization needs to analyze specific programs, the labor cost is high, and the effect is often not ideal

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Machine translation method, training method, corresponding device and electronic equipment
  • Machine translation method, training method, corresponding device and electronic equipment
  • Machine translation method, training method, corresponding device and electronic equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0050] The embodiment of the present application provides a machine translation method, such as image 3 As shown, the method includes:

[0051] Step S301: Obtain the semantic code corresponding to the source language input text.

[0052] In practical applications, the source language input text is generally a sentence input in the source language (that is, the smallest unit of translation processing), and the paragraphs, chapters, etc. input in the source language will be automatically split into sentences for processing.

[0053] In the embodiment of the present application, the semantic code corresponding to the source language input text is obtained, that is, the semantic code corresponding to the original input sentence is obtained directly, and then the obtained semantic code can be reused in step S302.

[0054] Step S302: Based on one or more target languages, decode the semantic codes respectively to obtain output texts in one or more target languages.

[0055] In th...

Embodiment 2

[0061] The embodiment of the present application provides a possible implementation manner as shown in the second embodiment on the basis of the first embodiment.

[0062] Specifically, since the semantic code corresponding to the source language input text acquired in step S301 needs to be reused, this embodiment of the present application may maintain a cache.

[0063] Then, after step S301, the semantic code can be stored in a preset cache, so that it can be used for subsequent decoding to different languages.

[0064] Furthermore, in step S302, for translation in any target language, the semantic code is obtained from the cache, and the semantic code is decoded based on the target language.

[0065] It can be understood that for the translation between two languages, it is only necessary to obtain the semantic code once from the cache and decode it based on the target language. For translation between multiple languages, semantic codes need to be fetched multiple times fr...

Embodiment 3

[0067] On the basis of Embodiment 1 or Embodiment 2, the embodiment of the present application provides a possible implementation as shown in Embodiment 3, wherein the machine translation method provided in the embodiment of the present application is applied to a machine translation network, the The machine translation network is still based on the design principles of encoder-decoder, such as Figure 4 As shown, including the pre-trained word vector space model, encoder and decoder.

[0068] Wherein, the word vector space model includes word vectors in multiple languages, and these word vectors in multiple languages ​​are based on semantic alignment. That is to say, the word vectors of each language are aligned in the same vector space. Compared with the relatively independent word vector spaces of different languages ​​in the prior art, the embodiment of the present application applies the word vector alignment to the translation task without the need for Other changes to ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a machine translation method, a training method, a corresponding device and electronic equipment. The machine translation method comprises the steps of obtaining a semantic codecorresponding to a source language input text; and based on the one or more target languages, respectively decoding the semantic codes to obtain output texts of the one or more target languages. According to the technical scheme provided by the invention, the semantic code corresponding to the source language input text only needs to be obtained once; therefore, the semantic codes can be repeatedly utilized, the output texts of different target languages are obtained through decoding based on the corresponding target languages, multiple times of calculation in the coding process are avoided,the overall calculation amount can be greatly reduced, and the processing efficiency of multi-language translation is effectively improved.

Description

technical field [0001] The present application relates to the technical field of natural language processing. Specifically, the present application relates to a machine translation method, a training method, a corresponding device and electronic equipment. Background technique [0002] Natural language processing is a technology for effective communication between humans and computers using natural language. Natural language is the crystallization of human wisdom, and the research on natural language processing is full of charm and challenges. After the development in recent years, the theoretical basis of natural language processing technology has become increasingly mature, and the scope of application has become wider and wider, which has driven a wave of industry boom. [0003] Counting from the 1950s, the research on natural language processing started from the machine translation system. Through a large number of scientific experiments, the public and the scientific c...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F40/58G06N3/08
CPCG06N3/08
Inventor 彭煦潭袁文博
Owner BEIJING SAMSUNG TELECOM R&D CENT
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products