JPH0258164A - Machine translation system - Google Patents

Abstract

PURPOSE:To post-edit already translated parts while translating untranslated parts by providing an interrupting means which interrupts the translation processing by an indication and a means which handles translated sentences and untranslated original sentences as documents independent of each other and dividing documents by an easy indication in accordance with progression of the translating work. CONSTITUTION:The interrupting means which interrupts the preliminarily designated translation processing in the stage, where original sentences stored in an original sentence storage part 5 are translated in a translation processing part 6 to store translated sentences in a translated sentence storage part 8, and a means 1 which automatically divides original sentences in the original sentence storage part 5, whose translation processing is interrupted with the interrupting means, in the position of translation interruption to handle divided them as two documents independent of each other are provided. Thus, the document is easily divided in accordance with the progression condition of the translating work to reduce the burden on an operator.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Field of Industrial Application) The present invention relates to a machine translation system that automatically divides documents.

(Prior Art) In recent years, machine translation systems using computers have been attracting attention. These machine translation systems are designed to perform translation processing using a document manually created using a keyboard or optical character recognition device, or a document already prepared on a computer or word processor.

Due to the ambiguity of natural language, it is extremely difficult to translate all documents accurately and appropriately. Therefore, current machine translation systems limit their processing targets to scientific and technical documents and manuals for various devices.

As trade has recently become more active, it has become necessary to translate a large number of manuals in conjunction with the import and export of products. A feature of machine translation systems is their ability to translate such large volumes of documents with stable quality.

In particular, word processors are widespread in Europe and America, and manuals are often stored on computer systems, so they can be easily used as the original text for machine translation systems.

Currently, these original documents are processed by operators according to specifications such as document length that can be handled by a machine translation system, and are translated, processed, etc. There is one.

That is, currently, the original document is divided and translated according to the specifications of the machine translation system used by the operator. However, depending on the division policy, the efficiency of translation and
This greatly improves the efficiency of document formatting after translation processing. For example, when dividing one document into longer lengths, either the effort required to divide the document is reduced, or more resources are required, such as document reading/writing time and the amount of memory used. Further, document formatting processing after translation processing tends to place a heavy burden on the system.

On the other hand, when one document is divided into shorter lengths, fewer resources are required when translating one document, but the burden on the operator for dividing the original document increases.

Furthermore, when resuming the translation of a document whose disciplinary action has been interrupted due to time constraints, there is a desire to edit the translation of the translated portion while translating the untranslated portion. However, with current systems this is not possible.

(Problems to be Solved by the Invention) As described above, conventional machine translation systems have the drawback that the translation efficiency when dividing an original document and the editing processing efficiency when restarting after translation interruption are not improved.

The present invention enables, when an original document to be translated is registered in a machine translation system, the task of dividing the original document in accordance with the specifications of the system and considering efficiency, to be performed during translation according to the wishes of the translator. It is an object of the present invention to provide a machine translation system that can reduce the burden on an operator, restart translation of a document whose translation has been interrupted, and edit the translated portion of the translated text.

[Structure of the Invention] (Means for solving the lesson problem) The present invention includes an original text storage unit that stores an input original text, and a translation unit that stores knowledge information used for translation processing of the original text stored in the storage unit. a dictionary section; a translation processing section that translates the original text stored in the storage section using knowledge information stored in the translation dictionary section; and stores a translated text obtained by translation processing in the translation processing section. In a machine translation system comprising a translation storage unit, a translation storage unit that stores , and an editing unit that performs editing processing on the translated text stored in the original text storage unit, the original text stored in the storage unit is processed in the translation processing. means for interrupting a pre-specified translation process in the process of performing translation processing in the translation unit and storing the translated text in the translated text storage unit; It has a means for automatically dividing the document and handling it as two separate documents.

Ru.

(Function) Conventionally, a large amount of translated documents has to be divided into parts by operators in advance according to the specifications of the machine translation system before translation processing, or large documents have to be translated directly, which places an unnecessary burden on the operator. However, by using this system, documents can be easily divided depending on the progress of translation work, making it possible to lighten the burden on the operator.

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

FIG. 1 is a block diagram of a Japanese-English mutual machine translation system according to an embodiment of the present invention. This system consists of a control unit 1 that controls the entire system, an input unit 2 that inputs instructions and text from the operator, and a display unit that displays the original and translated texts to be processed and information from the system. 3
, a document management section 4 that manages documents to be processed, an original text storage section 5 that stores documents managed by the document management section 4 or original texts obtained from the input section 2, and a translation of the original text stored in the original text storage section 5. A translation processing unit 6 for processing, a translation dictionary unit 7 that accommodates the translation processing unit 6 and knowledge information necessary for translation processing.
, a translated text storage unit 8 that stores the translated text obtained by the translation processing unit 6, and an editing unit 9 that edits the original text stored in the original text storage unit 5, the translated text stored in the translated text storage unit, or both. ing.

FIG. 2 is a diagram illustrating the manpower department 1 in detail.

This system basically consists of a human control section 21 that controls the input section, and a keyboard 22 that inputs data to be translated and instructions from an operator. Furthermore, a mouse 23, a character recognition device 24, a voice recognition device 25, a magnetic tape device 26, an optical disk device 27, etc. can be connected as necessary. It is also possible to input data from computer network 28.

FIG. 3 is a flowchart showing the operation sequence of the control section 1 that controls the entire system.

The control unit 1 performs input control (S31) and display control (S32).
, translation control (S33), document management control (S34), and other controls (S35), and monitors each component according to this operation sequence, and performs processing according to the information obtained therefrom. This information includes data indicating the status of each component, data to be sent to other components, and the like.

FIG. 4 explains step S31 in FIG. 3 in detail. In the input control, it is first checked whether there is any input from the input section 2. If there is no input at this time, the input control process ends.

(Proceed to step S32 in FIG. 3) If it is determined in step S41 that there is manpower, the process proceeds to step S42. S4
2) Check to see if it was done manually or if it was an instruction to start translation. If the input is a start instruction, the process moves to S43, and the translation processing section 6 is instructed to start translation. If it is determined in S42 that the instruction is not a start instruction, the process moves to S44. In S44, it is checked whether the input is an instruction to interrupt translation. If the input is an instruction to interrupt, the process moves to S45, and an instruction to interrupt the translation is given to the translation processing unit 6. If it is determined in S44 that there is no instruction to interrupt translation, the process moves to 546. In S46, it is checked whether it is an input or an instruction to divide a document.

If the input is an instruction to divide the document, the process moves to S47, and the document management section 4 is instructed to divide the document. If it is determined in step 546 that the instruction is not to divide the document, the process moves to step S48. 548
Now let's check whether human power is an editing command. If it is determined that the input is an editing command, the process moves to S49;
Editing instructions are given to the editing section 9. If it is determined in step 54g that the command is not an editing command (1111i), the process moves to step 54a.

S4a is S42. S44. S46. Interprets and executes instructions from operators other than S48.

FIG. 5 explains step S33 in FIG. 3 in detail. In this step, each processing section is controlled mainly in accordance with requests from the translation processing section 6. First, in S51, it is checked whether the translation processing unit 6 requests output to the display unit 3 or not. If there is no request, the process moves to S53. If there is an output request in S51, the output data is received from the translation processing section 6 and the data is passed to the display section 6. At this time, if necessary, it is also possible for the receiver to process the received data by adding data held by other components, such as the original text storage section 5 and the editing section 9, and to send it to the display section. When S52 ends, the process moves to 353. In S53, it is checked whether the state of the translation processing unit 6 has changed.

If there is no change, move to asss. If there is a change in status, information is received from the translation processing unit 6 and processing is performed. For example, it receives information such as when one sentence has been translated, when the entire specified document has been translated, when a paragraph break has been found, when a title has been found, etc., and the display section 6, document management section 4, etc. Send data to.

FIG. 6 is an operation sequence of the document management section 4 corresponding to step S34 in FIG. The document management section 4 first waits for an instruction from the control section 1 in step S61. When an instruction is received from the control unit 1, the process moves to step S62, and the receiver checks whether it is a received instruction or a reading instruction. In the case of a reading instruction, the process moves to step S63, and the instructed document data is sent through the control section 1 to the original text storage section 5, the translated text storage section 8, or both. If it is not a read instruction, the process moves to step S64, and the receiver checks whether the received instruction is a write instruction. In the case of a write instruction, the process moves to step S65, and the designated document data is received from the original text storage section 5, the translated text storage section 8, or both through the control section 1. If it is not a writing instruction, the process moves to step S66 and it is checked whether it is an instruction to divide the document. If the received instruction is to divide the document, the process moves to S67 and receives data necessary for dividing the document from the control unit 1. This data includes the number of characters translated up to the point of interruption. If the divided document is saved as a file, the file name is included, and if the document is sent to another user by e-mail, the destination address is included. Next step S6
8, the document data of the portion translated before the interruption is received from the original text storage section 5, the translated text storage section 8, or both through the control section 1. Next, the process moves to step S69, and the document data of the untranslated portion is received from the original text storage section 5 through the control section 1. Through this process, the document whose translation has been interrupted is automatically divided into translated documents and untranslated documents. Now, if the instruction received in S66 is not document division, the process moves to S6a and other processing is performed.

FIG. 7 is a flowchart showing a typical procedure for translating a certain document using a machine translation system. First, in step S71, a document to be translated is input. There are various input methods as shown in FIG. Next, the process moves to step S72, and pre-editing is performed. Pre-editing includes specifying parts that do not need to be translated, such as mathematical formulas and various symbols, and correcting errors when manually translated using the character recognition device 24. When the pre-editing is completed, the process moves to step S73 and translation is performed. At this time, the translation may be temporarily interrupted at some point in the operator's time and resumed later.

Further, the process may return to step S72 to correct the original text. Now, when the translation of the document is completed in this way, the process moves to step S74 and post-editing is performed. Post-editing involves correcting mistranslations, formatting, etc. Then step S
75 and output.

However, in the case of this conventional procedure, each document could not be post-edited until it had been translated several times.

FIG. 8 shows a point order in which the present invention is used to proceed with translation work while dividing a document in the middle of translation.

Steps S81 to S83 substantially correspond to S71 to S73 in FIG. What is different from FIG. 7 is step S84.
This is a division process. Here, the present invention is used to divide translated parts and untranslated parts into separate documents. As a result, untranslated portions are returned to step S82 or S83 to continue translation, while translated portions are moved to step S85 for post-editing work, improving work efficiency.

[Effects of the Invention] As described above, according to the present invention, when translating a large document, the document can be divided by simple instructions as the translation work progresses, and the translated portion can be divided while the untranslated portion is being translated. A machine translation system that allows for post-editing can be obtained.

[Brief explanation of the drawing]

Figure 1 is an overall configuration diagram of an embodiment of the system of the present invention. Figure 2 is a more detailed configuration diagram of the input section shown in Figure 1. Figure 3 is a control unit that controls the entire system. FIG. 4 is a flowchart showing step 531 in FIG. 3 in more detail.
The figure is a flowchart explaining step S33 in FIG. 3 in more detail, FIG. 6 is a flowchart showing the operation sequence of the document management unit 4 corresponding to step S34 in FIG. 3, and FIG. FIG. 8 is a flowchart showing a typical procedure for translation work using the present invention. 1... Control unit 2... Input unit 3... Display unit 4... Document management unit 5... Original text storage unit
6...Translation processing unit 7...Translation dictionary unit 8...
・Translation storage department・・Editorial department

Claims (1)

[Claims] An original text storage unit that stores an input original text; a translation dictionary unit that stores knowledge information used in translation processing; and a translation dictionary unit that stores knowledge information used in translation processing; a translation processing unit that translates the stored original text; a translation storage unit that stores the translated text obtained by translation by the translation processing unit; and a translation processing unit that stores the translated text translated by the translation processing unit; an editing unit that edits the translated text; an interrupting unit that interrupts the translation process by the translation processing unit according to an instruction; A machine translation system characterized by comprising means for handling it as a document.