JPH0244464A - Machine translation method - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] Industrial applications The present invention relates to a machine translation method.

BACKGROUND OF THE INVENTION Recently, machine translation systems have appeared that can automatically obtain a translated text simply by manually inputting the original text.

Such systems are aimed at translating human-generated original texts, and therefore little attention is paid to their overall control. That is, the conventional translation method only has a so-called single-task function of translating an input original text and outputting the translated text.

Problems that the invention aims to solveHowever, although the power of processors has improved and the O8 (operating system) has become capable of multitasking, machine translation systems that do this intentionally do not. There is not yet.

Means to solve the problem A machine translation mechanism is run in the background to execute multiple translation processes at the same time.

A plurality of documents can be translated simultaneously by batch processing in the background, and the translation processing speed per document can be increased.

Example An embodiment of the present invention will be described with reference to the drawings.

First, the concept or procedure of the machine translation system used in the present invention will be explained with reference to the algorithm shown in FIG. 3 based on a general syntax transfer method.

This is, for example, an English-Japanese machine translation system,
After reviewing the original text, we selectively perform morphological analysis for spell checking. Then, a morphological analysis is performed. This is a word-by-word analysis, and is an analysis process that focuses on special characters such as hyphens. Next, parse the syntax. This is syntactic analysis to understand the structure of a sentence. From this syntactic analysis result, an English structure identification process is performed, such as eliminating semantically incorrect analysis, to create an English intermediate structure, and then a syntactic conversion process is performed. That is, an intermediate structure based on English phrases is converted into an intermediate structure that easily represents Japanese phrases. Then, generate Japanese. That is, in order to generate natural Japanese sentences from Japanese intermediate structures, syntax is first generated, and then morphemes are generated to arrange the order of particles and auxiliary verbs. After this, the translated text, for example, Japanese, is finally output.

Here, the input original text is subjected to morphological analysis, and the results are fed back to the user for pre-editing, or the translation results are displayed in bilingual terms and used for post-editing to make them what the user desires. - Has a post-editing function.

In this embodiment, such translation processing as a whole is considered as one machine translation mechanism, and this is efficiently operated in a batch system. Note that the actual translation is not affected by the translation method shown above.

Further, the block diagram configuration is as shown in FIG. 2. That is, a second translation section 1 that executes translation, a document list management section 2, a first translation section 3 that executes translation processing to be described later, and an input section 4 that performs manual command and data reading.
It consists of an output section 5 that outputs translation results, etc., and a control section 5 that controls these as a whole.

In such a configuration, the translation control method according to this embodiment will be explained with reference to the flowchart of FIG. first,
Conventionally, a document that has been processed as necessary is passed to a machine translation mechanism, and the document is edited after waiting for the translation process to finish. In this embodiment, in addition to such a conventional method, a batch processing section as shown in FIG. 1(a) is prepared as a background.

That is, the foreground processing unit of the first translation unit 3 shown in FIG. 1(b) prepares a document list file containing a list of documents to be batch translated, and the batch processing unit shown in FIG. The translation is performed one by one starting from the first document in this document list file.

Then, it is necessary to perform processing such as removing or marking documents for which successive translation has been completed from the document list file.

The batch processing shown in FIG. 1(a) is activated by selecting the background execution function in FIG. 1(b). Also, it is forcibly terminated by an interrupt. The translation results obtained by such batch processing are also equivalent to conventional translation results, and therefore can be post-edited.

Therefore, according to the method of this embodiment, it is possible to translate a plurality of documents at the same time, and the translation processing speed per document is dramatically improved.

In this example, only one translation is executed in the background, but it is possible to execute multiple translations depending on the processor's ability and usage status. Furthermore, by prioritizing documents to be batch processed, it is possible to have the documents translated in an order according to need, rather than on a first-come, first-served basis.

Furthermore, by providing a mechanism to interrupt batch translation, it is possible to carry out a large number of translations over several days almost continuously, interspersed with stopping the machine as needed. By preparing a plurality of batch processing sections shown in ) and the corresponding processing sections shown in FIG.

Effects of the Invention In the present invention, as described above, the machine translation mechanism is run in the background to execute multiple translation processes at the same time, so multiple documents can be translated simultaneously through batch processing in the background. , the translation processing speed per document can be dramatically improved, and the user interface can also be improved.

- Fig. 2 is a block diagram, and Fig. 3 is a flowchart showing the syntax transfer method.

Co., Ltd.

[Brief explanation of the drawing]

The drawings show an embodiment of the present invention, and FIG. 1(a) is a flowchart of the background batch processing section, and FIG. 1(b) is a flowchart of the background batch processing section.
are the flowcharts of the foreground processing section (a) (b)

Claims (1)

[Claims] A machine translation method using a machine translation mechanism, characterized in that the machine translation mechanism is run in the background to execute a plurality of translation processes simultaneously.