JPH06124303A - Machine translation system - Google Patents

Abstract

PURPOSE:To rearrange respective divided and translated elements in the machine translation system by giving numbers corresponding to divided elements of an original sentence to the respective divided and translated elements and only indicating given numbers. CONSTITUTION:When the original sentence is long at the time of translation, the original sentence is divided at a dividable position and numbers are given to the respective elements. Translation results are also given corresponding numbers and a translator only indicates the number of an element to be rearranged, so that the divided translation results can be rearranged in desired order.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine translation system for translating a translated sentence in a first language into a translated sentence in a second language.

[0002]

2. Description of the Related Art Conventionally, machine translation systems are good at translating short and concise translated sentences, and it is difficult to perfectly and accurately translate long and complicated translated sentences. Therefore, in order to obtain a more accurate translation result, the translator has to divide a long translated sentence as a pre-editing work. Or, in order to reduce the burden on such translators, if the input translated text is longer than a certain length set by the system, treat it as an independent sentence or phrase even if it is divided. A method has been devised in which the system determines the translated sentence at a position where the translation can be performed and divides the translated sentence, and performs translation processing individually.

For example, in Japanese Unexamined Patent Publication No. 1-273171, after the system divides a translated sentence, the system judges and supplements a word connecting a preceding sentence and a succeeding sentence, and separates each divided sentence. A method of completing the sentence as an independent sentence is proposed. In addition, JP-A-2-59860 and JP-A-1
In Japanese Patent Publication No. 152573/1989, individual elements that have been divided and translated can be rearranged according to instructions from a translator. By doing so, it is possible to prevent the whole translated sentence from collapsing due to the fact that the translated sentence is long and the constituent elements of the sentence cannot be put together or the recognition of the modification destination fails. The result translated as a partial translation is
The translator may freely rearrange or connect by editing means to compose the entire sentence. Since the partial translation itself is likely to be translated correctly, it is possible to save the trouble of correcting the translation result.

[0004]

However, in the conventional method, the system determines the order of division, rearrangement after translation, and order of concatenation, or it is left to the translator's editing work. When the system does, the result may be contrary to the translator's intention, and when the translator rearranges, refer to the translated sentence and specify the partial translation in the order you want to rearrange by specifying the range etc. I have to tell you.
The work of actually rearranging takes time and effort.

The present invention has been made in order to solve the above-mentioned problems, and prevents the translated sentence from collapsing due to a long translated sentence, and divides the translator to obtain a desired translated sentence. An object of the present invention is to provide a machine translation system that can reduce the trouble of rearranging the created elements.

[0006]

To achieve this object, a machine translation system of the present invention provides an input unit for inputting a translated sentence in a first language, and a translated sentence input by the input unit. A dictionary that stores information necessary for translating into a translated sentence in two languages, a unit that translates a translated sentence input by the input unit using the dictionary, and an output that outputs a translation result by the translation unit. And means,
When the translated text input by the input means is long,
The translated sentence is divided at division positions where it can be processed as an independent sentence or phrase, each divided element is given a code that does not overlap, then translated, and the corresponding code is also given to the translation result. It is characterized in that a translation result sentence is generated by rearranging each element according to this code.

[0007]

The machine translation system of the present invention having the above-mentioned configuration uses a dictionary that stores information necessary for translating a translated sentence in the first language input from the input means into a translated sentence in the second language. And translate it into a translated sentence. When the translated sentence is long, the translated sentence is divided at division positions where it can be processed as an independent sentence or phrase, and each divided element is given a unique number and then translated A translation result sentence is generated by assigning a corresponding number to the output and rearranging the numbers.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an outline of the machine translation system of this embodiment.

In FIG. 1, the machine translation system 1 is
A data processing unit 10 including a computer is connected to an input unit 11 including a keyboard, a display unit 17 including a CRT, and a printing unit 18 including a scanning printer. . Further, inside the data processing unit 10, a translated sentence and a translated sentence file are stored by using an auxiliary storage unit such as an editing unit 12, a translation unit 13, a hard disk, an optical disk, etc., which is realized by software using a function of a computer. Part 14 (hereinafter referred to as file storage unit), dictionary 15 having information necessary for translation, partial number management for managing the correspondence between the numbering process of each element instructed to be divided by the editing unit 12 and the translation result. The unit 16 is provided.

Next, the operation of the machine translation system configured as described above will be described in detail by taking a system for translating from Japanese to English as an example.

FIG. 2 is a flow chart schematically showing the structure of the program until the input translated sentence in the first language is converted into the translated sentence in the second language and is completed as a translated sentence.

The translated sentence in the first language input from the input unit 11 (S20) is edited by the editing unit 12 as necessary for typographical errors and omissions (S21), and then a sentence is stored in the file storage unit 14. Each number is sequentially assigned with a non-overlapping number and then temporarily stored.

When the translated sentence to be translated is ready and a translation execution instruction is given from the input unit 11, the translated sentence is sent to the translation unit 13 (S23). The translated sentence is first divided into translation units using punctuation marks (“.”) And the like (S
23a), the morpheme analysis process divides each word into translation units using the dictionary 16 (S23b), and the analysis process performs syntactic analysis and semantic analysis based on the dictionary information of the first language (S23c). Next, in the conversion process, the analyzed translated sentence is converted from the first language into the translated sentence of the second language (S23d). The translated sentence converted in the conversion process is generated in the correct second language syntax based on the dictionary information of the second language in the dictionary 16 by the generation process (S23).
e) Further, by the morpheme generation process, the word form change process and the like are performed to complete the process (S23f). The morpheme analysis, analysis, conversion, generation, and morpheme generation in the translation unit 13 are all performed using the dictionary information stored in the dictionary 15.

The translated sentence thus obtained is assigned the same number as the corresponding translated sentence and stored in the file storage unit 14 (S24). That is, one sentence number is assigned to a pair of the translated sentence and the translated sentence.

3 and 4 show an example of division processing and rearrangement by number designation when the translated sentence is long. With respect to this example, a flowchart showing the steps of the process of FIG. 5 will be described. When the translated sentence is long, the translator uses the editing unit 12 to instruct the division position so that the system can easily translate the sentence (FIG. 3A). The partial translation management unit 16 divides the translated sentence according to the instruction (S5
1), numbering each element (S52) (FIG. 3)
(B)). The translator edits the beginning and end of each divided element into an appropriate form (Fig. 3 (c)) and translates it (S
53). When the translation result is obtained, the translator indicates the sorting order by number (FIG. 4 (a)). The system rearranges the translated sentences according to this instruction (S54) (FIG. 4).
(B)), the translator adjusts the connection portion to complete the translated sentence (FIG. 4 (c)).

In this example, the translator instructs the division position of a long translated sentence.
The position that can be divided, for example, the position where the connective or the connective particle appears, the continuous discontinuation form, the adnominal form, etc. may be stored and the division may be performed by the system. The system can also adjust the beginning and end of sentences after division or connection. Furthermore, in this example, an example is shown in which the screens for Japanese and English are separate, but the parallel translation screen is more convenient for rearranging.

[0018]

As is clear from the above description, according to the machine translation system of the present invention, when a translated sentence is long, it is translated after being divided and translated at a position where it can be processed as an independent sentence or phrase. To sort the results, all you have to do is specify the number, which reduces the effort of completing the translation.

[Brief description of drawings]

FIG. 1 is a block diagram showing an outline of a machine translation system of the present invention.

FIG. 2 is a flowchart schematically showing a configuration of a program for converting a translated sentence into a translated sentence.

FIG. 3 is a diagram showing an example of a process of dividing a translated sentence.

FIG. 4 is a diagram showing a processing example of rearranging each element as a result of translating a divided translated sentence.

FIG. 5 is a flowchart of a process of dividing a long translated sentence and a process of rearranging by specifying a number.

[Explanation of symbols]

 1 Machine Translation System 11 Input Section 13 Translation Section 15 Dictionary 16 Partial Translation Management Section 17 Display Section

Claims (1)

[Claims] 1. An input unit for inputting a translated sentence in a first language, and a dictionary for storing information necessary for translating the translated sentence input by the input unit into a translated sentence in a second language, In a machine translation system comprising a translation means for translating a translated sentence input by the input means using the dictionary, and an output means for outputting a translation result by the translation means, a translated text input by the input means is provided. If the sentence is long,
The translated sentence is divided at division positions where it can be processed as an independent sentence or phrase, each divided element is given a code that does not overlap, then translated, and the corresponding code is also given to the translation result. The machine translation system is characterized in that a translation result sentence is generated by rearranging each element according to this code.