JPH06119373A - Machine translation system - Google Patents

Abstract

PURPOSE:To provide a translation system in which it can be distinctly known where a problem exists in the case that an inappropriate translation is outputted and a user can easily judge whether an individual rule inside a system should be replenished or what translation dictionary should be generated, corrected or deleted in order to obtain the translation of good quality. CONSTITUTION:A dictionary part 14, an individual rule part prepared separately inside the system in order to output special translation incapable of being dealt with by the dictionary part 14, a translating part 13 for translating first language into second language by using the dictionary part 14 and the individual rule part, a post-editing part 15 for editing the second language of a translated result, and a display part 16 for displaying the first language and the second language are provided, and further in detail, the display part 16 is provided with a display control means 15a for displaying whether an individual equivalent is based on the individual rule or the translation dictionary part 14 and the name of the translation dictionary clearly to the user.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine translation device for performing translation processing using a plurality of translation dictionaries.

[0002]

2. Description of the Related Art Recently, a system for automatically translating an input original sentence using a computer has been receiving attention. Basically, such a machine translation device morphologically analyzes the input source text,
Divide into predetermined processing units such as words (phrases) by performing syntax analysis, search multiple translation dictionaries for each processing unit,
It is configured to obtain a translated sentence by finding a corresponding translated word or the like and combining them according to a predetermined translated sentence generation rule. The translation dictionary of such a machine translation device is generally used to output a dictionary storing a part of speech and a translated word corresponding to a basic headword (hereinafter referred to as a basic dictionary) and a translation desired by an individual user. It is provided with a plurality of translation dictionaries such as a dictionary constructed by the user (hereinafter referred to as a user dictionary) and a dictionary (hereinafter referred to as a specialized dictionary) in which parts of speech and translations are stored in correspondence with technical terms. As a method of searching for these multiple translation dictionaries, conventionally, in addition to the method of searching the user dictionary, the specialized dictionary, and the basic dictionary in order, a method of searching based on the priority order among the multiple translation dictionaries has also been proposed. There is.

However, which translation dictionary is actually searched is determined by whether or not the entry word is registered in each translation dictionary, and in addition, a special translation that cannot be handled by the dictionary is provided inside the system. Since it depends on whether or not there are individually prepared rules, it is often impossible for the user to judge. Also, a system has been proposed to indicate which dictionary the source is from, but if a special rule is prepared inside the system for words that cannot be supported by the dictionary, one of the above dictionary, or I don't know if it's due to a special rule inside the system. For example, the verb "Iru" is used as "(Nominative) is in (Place)." In an ordinary dictionary, the English verb "be" is often applied.

However, depending on the input sentence, there i
In some cases, the s syntax is more appropriate. In this way, when the sentence structure needs to be translated according to the relationship of the words before and after it, it is not possible to keep a translation in the dictionary, so a special rule is prepared and the system applies that rule. However, the user had no way to determine if it came from the dictionary or if it was due to the special rules above.

[0005]

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and displays in a user-friendly manner whether the translation is based on individual rules inside the system or which translation dictionary. By doing so, it is possible to clearly understand where the problem occurs when an inappropriate translation is output, and whether individual rules inside the system must be enhanced in order to obtain a good translation, or which translation dictionary is created Another object of the present invention is to provide a translation device that allows the user to easily determine whether it is better to correct or delete.

[0006]

In order to achieve this object, a machine translation apparatus of the present invention has an input unit for inputting a first language and a form suitable for machine translation of the input first language. A pre-editing section for editing, a first dictionary storing parts of speech corresponding to headwords in the first language and translations in the second language, a second dictionary constructed by each user, and technical terms stored Using the dictionary unit having the third dictionary described above, the individual rule unit individually prepared in the system for producing a special translation that cannot be handled by the dictionary unit, and the dictionary unit and the individual rule unit. A translation unit for translating the first language into the second language, a post-editing unit for editing the second language that is the translation result, and a display unit for displaying the first language and the second language , And more specifically, the display unit displays individual translated words And a display control means for displaying clarity whether by separate rules, or whether by how translation dictionary of the dictionary portion to the user.

[0007]

In the present invention having the above-mentioned structure, the first language is input to the input unit, and the input first language is preedited into a form suitable for machine translation, and then the dictionary is referred to by the translation unit. Is translated into the second language, the second language as the translation result is edited in the post-editing section, and the first language and the second language are displayed on the display section. The display unit displays on the display unit in a user-friendly manner whether each output translated word is based on an individual rule or which translation dictionary of the dictionary unit.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment embodying this embodiment will be described below with reference to the drawings.

Here, an embodiment in which the first language is Japanese and the second language is English will be described.

FIG. 1 is a block diagram functionally showing the hardware configuration of a machine translation device according to this embodiment. As shown in FIG. 1, the machine translation apparatus 1 includes a data processing unit 10 including a computer, a keyboard 11 as an input unit, a CRT 16 as a display unit, a scanning printer 17 as a printing unit, and the like. It consists of

In the data processing unit 10, the pre-editing unit 12 for editing the input sentence input to the input unit 11 into an appropriate form as needed, and receiving the input sentence from the pre-editing unit 12 receive the first sentence. A translation unit 13 for translating into two languages, a dictionary unit 14 stored in a hard disk or an optical disk, a translation unit 1
The post-editing unit 15 for editing the translation result sentence obtained in 3 is realized as software and is incorporated.

FIG. 2 is a diagram showing the structure of the dictionary unit 14. The dictionary unit 14 includes a basic dictionary 20 and a user dictionary 22.
And a specialized dictionary 24.

FIG. 3 shows a flow chart used for explaining the operation of this embodiment.

First, when a Japanese sentence which is an input sentence, for example, "There is an excellent engineer at a local factory." Is input from the input unit 11 (S1), the Japanese sentence is input by the pre-editing unit 12 as necessary. After being edited into an appropriate form (S2), it is sent to the translation unit 13. The Japanese sentence sent to the translation unit 13 is morphologically analyzed by referring to the dictionary unit 14 in the morphological analysis unit 13a as shown in FIG. 4, divided into words, and word information or the like is added to each word (S3). ). Further, as shown in FIG. 5, in the syntactic analysis unit 13b, the words are collected into phrases while similarly referring to the dictionary unit 14, and syntactic analysis such as dependency analysis is performed (S4). After that, in the syntax conversion unit 13c, Japanese is converted into an English structure based on the result of the syntax analysis as shown in FIG. 6 (S5). That is, here, each case element of Japanese is associated with an English case element, and transformed into an appropriate English structure.

The sentence converted into the English structure by the syntax conversion unit 13c is delivered to the syntax generation unit 13d. First, the dictionary unit 14 is similarly searched for each word or group of words. Next, based on the translated word generation rule, the correct syntax is generated as English (S
6). Subsequently, the morpheme generating unit 13e performs morpheme generation such as word form change (S7). The English translation generated in this way is freely edited by the post-editing unit 15 (S
8) is displayed on the display unit 16 or is output by the printing unit 17 (S9).

In the system configured as described above, the display method on the display unit 16 and the process up to that point will be described in detail.

In the morphological analysis unit 13a, a plurality of dictionaries of the dictionary unit 14 are searched to divide the input sentence into words, but if the divided words at that time are present in the user dictionary, they can be drawn from the user dictionary. An ID number (0 in this case) indicating a word is assigned to a Japanese entry word, and if it exists in the specialized dictionary, an ID number (2 in this case) indicating a word that can be drawn from the specialized dictionary is used in Japanese. If it exists in the basic dictionary, the ID number (1 in this case) indicating that it is a word that can be drawn from the basic dictionary is assigned to the Japanese entry word. After that, in the syntax analysis unit 13b,
The relevant dictionary is searched, and necessary word information is added to each word, and at the same time, the dependency is compiled.

As shown in FIG. 7, in the syntax conversion section,
The dictionary unit 14 is searched for each entry word of the Japanese sentence after the syntax analysis. I indicating that it is a user dictionary
If the D number is 0, the user dictionary is searched and the translation registered therein is applied, and if the ID number 2 indicating that it is a specialized dictionary is attached, the specialized dictionary is searched and the translation is applied. If there is an ID number 1 that indicates that the dictionary is a basic dictionary, the translation that is registered in the basic dictionary can be applied, but that translation is not applied. If you want to apply the translation of the prepared rule, set a flag indicating that the individual rule is prepared in the dictionary. (here
FLAG = T) Even if the word with this flag has an ID number 1 indicating that it is a basic dictionary, the translation of the basic dictionary is called instead of applying the translation of the basic dictionary. Throughout the process of translation, a Japanese headword, an ID number, and a flag indicating that there are individual rules are always held.

As shown in FIG. 8, when displaying the translation result, the display control unit 15a searches the generated English for the ID number and the flag attached to the Japanese entry word based on the designation by the user. Then, if the ID number is 0, it is "user dictionary", if the ID number is 1, it is "basic dictionary", I
If the D number is 2, "special dictionary" is displayed to the user, and if it is output from the individual rule, "individual rule" is displayed to the user. If the translation of the dictionary is not appropriate by looking at the result displayed in this way, the corresponding dictionary can be corrected or a necessary dictionary can be created. In addition, there is an advantage that the dictionary can be improved smoothly, such as requesting individual rules to the system developer if it is judged that the dictionary cannot provide an appropriate translation.

The present invention is not limited to the above-mentioned embodiment, and other methods of displaying it to the user in an easy-to-understand manner are conceivable. Also, if the user wants to search only the words output from the user dictionary, "user dictionary"
If you specify, all the words output from the user dictionary can be highlighted.

By using this embodiment, the user can clearly understand where the problem is in the portion where the output result is inappropriate. Thereby, the quality of each dictionary including the user dictionary can be further improved.

[0022]

As is apparent from the above description,
According to the machine translation device of the present invention, it is possible to easily maintain and improve the dictionary by clearly indicating to the user which dictionary and the individual rule the output result is, and it is expected that the quality of the translated text is improved. .

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a machine translation device to which this embodiment is applied.

FIG. 2 is a diagram showing a configuration of a dictionary used in this embodiment.

FIG. 3 is a flowchart of the entire translation process provided for explaining the operation of the present embodiment.

FIG. 4 is a flowchart of a morpheme generation process used to explain the operation of this embodiment.

FIG. 5 is a flowchart of a parsing process used for explaining the operation of the present embodiment.

FIG. 6 is a flowchart of a syntax conversion process used to explain the operation of this embodiment.

FIG. 7 is a diagram showing a search result of a dictionary used for explaining the operation of the present embodiment.

FIG. 8 is a diagram showing an example of a post-editing screen obtained as a result of the action of the present embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Machine translation device 10 Data processing section 11 Input section 12 Pre-editing section 13 Translation section 14 Dictionary section 15 Post-editing section 15a Display control section 16 Display section 17 Printing section

Claims (1)

[Claims] 1. An input unit for inputting a first language, a first dictionary storing a part of speech corresponding to a headword of a basic word of the first language, a translated word of the second language, and the like, and Second dictionary containing user-created dictionaries or technical terms
Using the dictionary unit including the dictionary, the individual rule unit for producing a special translation that cannot be handled by the dictionary unit, and the first and second dictionary units and the individual rule unit, In a machine translation device comprising a translation unit for translating into a second language, a post-editing unit for editing a second language as a translation result, and a display unit for displaying the first language and the second language. The display unit includes display control means for distinguishing and displaying whether each output translated word is based on an individual rule, or whether the translated word is based on one of the translations of the dictionary unit. Machine translation device.