JPH02144767A - Mechanical translation system - Google Patents

Abstract

PURPOSE:To simplify the editing operation of an operator and to easily obtain the proper translated sentences by producing automatically the substitute candidates based on the attribute information on the translated words and then carrying out the substitute. CONSTITUTION:When an original sentence is inputted by an operator via an input part 1, the corresponding character codes are set to an input buffer in an editing control part 4 and displayed at a display part 8. Then the operator finishes the input of one sentence and pushes a translation instruction key. Thus the part 4 instructs the start of a translation process to a translation part 5. The part 5 uses an affixial word connection/inflection dictionary 6a contained in a translation dictionary 6 to convert the words into their original forms and then retrieves a word dictionary 6b to obtain the parts of speech and the translated words. Then the structure of a Japanese sentence is produced based on an analysis grammar 6c, and the Japanese word structure is converted into the English word structure based on a conversion grammar 6e. The word order of an English sentence is decided by a production grammar 6f and converted into the word strings. Then the word forms, etc., are changed by a morpheme production grammar 6g for production of a translated sentence.

Description

DETAILED DESCRIPTION OF THE INVENTION [Aim of the invention] [Field of industrial application] Akira Motoya relates to an Iso-mechanical abstraction system that can efficiently perform editing work after machine translation.

(Prior Art) In recent years, a system that uses a combiere to machine-translate an input original text has been attracting attention. This machine-reduced translation system basically divides the input original text into predetermined processes such as morphological analysis and grammatical analysis, and uses a translation dictionary for each processing unit. Do you respond immediately? It is configured to find a translated word (translated phrase), etc., and combine it with the translated word according to a predetermined translation rule to obtain the translated text.

However, since semantic interpretation and context understanding techniques for natural languages have not yet been sufficiently established, it is still very difficult to reliably obtain appropriate translations. However, there is still a sense of security that the operator can edit the translated text appropriately.

Additionally, in order to support editing work to obtain appropriate translations, a function called translation word selection has conventionally been provided. With this function, when an operator wants to translate a certain word in a translated text into a different translation, by specifying that word or the corresponding word in the original text, the system displays the contents of the specified QLM dictionary. Then, the operator looks at the displayed content and selects one of a plurality of translated words, and the system automatically changes the translated word.

However, in order to obtain an appropriate translation, it is not enough that the translation word is appropriate.

The editing effect of appropriate tenses, states, preverbs, articles, etc. is great. For example, when translating the sentence ``The program refers to.'' into English, ``The program refers to [something e].'' There are two possible interpretations: ``(something) refers to the program.'' and either interpretation is appropriate depending on the context.If the second interpretation is appropriate, then the 14th interpretation is applied, r A pro
If you get the translation ``A program is consult
It can be made more appropriate by replacing the verb in the active voice with the passive voice, such as -edJ.

Conventionally, such editing work has been performed by an operator moving a cursor to an appropriate location and deleting or inputting a single character.

Performing such li editing work repeatedly interrupts the flow of thoughts of the operator, increases psychological burden, and
The problem was that it was easy to cause input errors and other operational errors.

(The problem that Seimei is trying to solve) As described above, conventional machine translation systems have the drawback of not being able to efficiently perform editing work by optimizing tenses, states, prepositions, articles, etc.

This letter was written with these circumstances in mind, and the purpose of this letter is to avoid the long-distance conversion work that is done depending on the attributes of translated words during editing, and to create a translation dictionary. One object of the present invention is to provide a machine translation system that can simplify the editing process of an operator by using the attribute information of the stored translations U(j).

[Katsumei composition]

(Means for solving the problem) In order to achieve the above purpose, the present invention provides a means for instructing the correct target translation nfJ (phrase) from the input section, and a translation knowledge m,...
I；fwζ This translated word (phrase) is contained
What is required is to have means for generating primary replacement candidates from this, means for displaying the candidates, and means for an operator to selectively instruct the candidates. i-Characterized.

(Operation) In this method, for a translation M (phrase) for which an operator has instructed to correct, the system automatically creates replacement candidates based on the attribute information of the translation MI (phrase) stored in the translation knowledge dictionary. and then replaces the Houpu selected by the operator with the original translation.

(Embodiment) FIG. 1 is an overall block diagram of a Japanese-English machine translation system as an embodiment of the present invention. This system consists of an input section 1 consisting of a keyboard, a JM, an original text storage section 2 that stores Japanese d5c inputted as a sentence, and a translation section 1 that stores 31 English sentences corresponding to the input Japanese sentences as translated sentences. A translation control unit 4 that controls the entire system, a translation unit 51 that executes the #fl translation process, and a translation dictionary 1FG that accommodates the knowledge J11 used in the translation process.

It is composed of a display control section 7 for controlling the display of the original text, translated text, etc., a display section 8, and a printing section 9 for performing RFF8 on the original text, translated text, etc. and outputting it as a copy.

FIG. 2 shows a row of key layouts of the input unit 1. As shown in FIG.

Input section 1 has the following various keys in addition to the letters for English language skills: IJf! It is growing.

Translation instruction key: 100 Edit key: 101-106 Function key: 110-121 Cursor key: 130-133 Key of the pond FIG. 3 is a diagram showing an example of the layout of the display section 80 drawing. The input original text is displayed in the original text display area on the left side of the screen. The translated text obtained as a result of the translation process is displayed in the translated text display area on the right side of the screen at a position corresponding to the original text. Additionally, the editing area at the top of the screen can be used to display information necessary for various types of editing.

FIG. 4 is a flowchart showing an example of the flow of dialogue translation processing by the editing control unit 4. In dialog translation processing, the ilA concentration control leg 4 step 5L-8 allows the operator to input the original text and edit the corresponding translated text as appropriate.
6, it is monitored whether there is any key input from the input unit 1 or whether a translation completion signal is received from the translation unit 5, and if there is any key input, 1 performs the corresponding process.

If the operator operates any of the function keys, this is detected in step S5, and the process corresponding to that function key is performed in step 812.

If the operator operates any of the editing keys, this is detected in step S4, and the process corresponding to the editing key is performed in step 811.

When the operator operates the cursor key 130 or other control keys, processing for moving the cursor corresponding to the key operation and processing for the cursor are performed in Step 7S13 via steps 81 to S6.

When the operator inputs the original text by operating the character keys, each character key is detected in step S3.

Corresponding character codes are sequentially set in input buffers (not shown) in the TA collection section 4 (step S8).

This input original text is displayed in the original text display area of the display unit 8 via the display control unit 7 (step S9). If the operator presses the translation instruction key 100 at any time when the manual input of one sentence is completed, this key is detected in step S2, and the 1-editing phloem 4 inputs the original text in the buffer into the translation section 5I. is supplied to instruct the start of translation processing (step 5IO). If input editing such as correction, insertion, deletion, etc. is required during the original text, move the cursor to the desired editing location using the cursor keys, and then press the insert key 10.
1. This can be done using an edit key such as delete key 102.

FIG. 5 is a flowchart showing the flow of translation processing in the translation section 5.

The form'A analysis section (S51) analyzes the human-powered original text.

QIs with changes in word endings, etc. using the Old Number Change Dictionary 6a.
Convert AH to its original form (basic form).

The dictionary search unit (S52) searches the word/idiom dictionary 6b for each hymn that constitutes the input original text, and obtains information such as its part of speech and translation.

The analysis control unit (S53) performs a process of sending translation candidates to the syntax analysis unit.

The syntax analysis unit (S54) uses the analysis grammar 6C to analyze the syntax of the translated sentence, and generates the structure of a Japanese sentence. If the analysis fails, the process returns to the analysis control section.

The structure conversion unit (S55) converts the Japanese structure into the English structure using the conversion grammar 6d. If rejected,
Return to the parser.

The syntax generation unit (S56) determines the English word order from the English structure according to the generative grammar 6e and converts it into a word string.

Shape I2! The l-element generation unit (S57) uses the morpheme generation grammar 6f
The translated sentence is completed by changing the word form etc.

Each of the above-mentioned processing units outputs a signal to the editing system section 4 at the time of starting processing.

Pick up the number 4 again, 1i! Cape Kasei part 4 is step S
When the translation completion signal is detected in step S7, the translation candidates and auxiliary information obtained from the translation unit 5 are sent to the display unit 7 to display the translated text (step 87).
. The display control unit 7 places the translation candidate in the translation display area at a position [11
In addition to displaying ``i''information,; 1n supplementary information is displayed,
The translated word is displayed with high brightness or reversed so that the operator can identify it.

In this state, when an edit key is pressed, the 2 edit 1 interpolation unit 4 detects this in step S4, and performs the ia processing corresponding to each key on the nl or phrase indicated by the cursor at this time. Execute [Step S11].

For example, by operating the insert key 101, a character is inserted before the cursor is placed. By operating the delete key 102, the character string in the range indicated by the cursor is deleted. According to the operation number of the movement key 103, the range indicated by the cursor 41 is moved. By operating the cancel key 104, the effects of the keys 101, 102 and 103 are nullified. Modifications are made by operating the key 105 and using auxiliary information to select δ!, which is indicated by the cursor. Other dependencies in the phrase display the housode.

The cursor keys include a cursor movement key 130 to move the cursor in each direction, a unit switching key 131 to change the unit in which the cursor moves, and an inter-area movement key 132 to move the cursor between display areas.
It consists of enlargement/reduction keys, etc. for reducing the size of one cursor on a character-by-character basis or enlarging on a word-by-word basis.

Therefore, the operator can interactively obtain a translated text while inputting the original text and editing the translated text as appropriate.

Furthermore, the following functions can be realized using the mechanical key.

From the operation button of the translation display key 110, the translation of the word in the translation can be displayed. By operating the dictionary display key 111, a dictionary can be displayed for a word or phrase in the original text. New words and phrases can be registered by operating the dictionary V registration key 112. By operating the dictionary deletion key 113, words and words registered through dictionary registration can be deleted. By operating the partial translation key 114, the partial translation is displayed when translation fails.

The function keys 115 to 121 are related to the present invention and can realize the following functions.

All the times that can be created by changing the tense of the specified translation verb by operating the function key 115.
ll conversion replacement candidates can be displayed.

By operating the function key 116, it is possible to display summer clothing replacement candidates that are created by changing the verb to the particle of the designated translated word.

By operating the function key 117, it is possible to display all possible tense/transformation substitution candidates ttll that can be created by changing the tense or deteriorating the verb of the specified translated verb.

By operating the function key 11B, it is possible to refer to the first literal word in front of the specified translation 1411, which can be replaced with a word (associative disguise replacement candidate).

1:! With one operation of the li key 119, it is possible to display article conversion replacement candidates that can be substituted for the article of the specified translated word.

By operating the function key 120, it is possible to display singular/plural conversion candidates 6n that can be substituted for the noun of the translated word specified by the operation of the function key 120.

In addition, by operating the function key 121 ζ, it is possible to determine the part of speech of the specified translated word and display all possible replacement candidates according to each part of speech.

In the system configured as above, function key 1
15 to 121 will be described in more detail. FIG. 6 is a diagram showing the flow of tense conversion and replacement processing. First, the operator looks at the translated text displayed on the display section 8 and moves the cursor to the word E for which the translation is to be changed. Check the translation information contained in the translation results sent to you and check whether the translated word is a verb. Check to determine whether tense f conversion is possible (603). The translation information is the attribute information retrieved from the translation dictionary by the dictionary search unit (852) and the translation information in the translation intelligence process. This is attribute information attached to.

If conversion is possible, 1i! Collection part 4 uses attribute information such as '1lil1' for the specified translated word to change it to other tenses, and uses the translated word as is and II! Generate all translated words that can be converted into L (604) Display them in the percentage display section 8 (605)
). When one of the replacement candidates is selected by the operator (
606), the editing system misaki department 4 changes the translation and displays the result (607).

The operator's operations will be described below with reference to a diagram showing a display screen in the tense conversion process shown in FIG.

``He wrote a book.'' Give an example of a translation of the original text and explain. First, the original text and the corresponding translated text are displayed (FIG. 7(a)). If the operator wants to change the tense of "wrote" after seeing this display, move the cursor to 'wrote' and press the 1 tense change 1 back key 11
When you press 5, the tense and other attributes of ``wrote'' are checked, and ``wrote'' is changed to a tense other than past tense.''
writes', "1a wrlt%ing", "h
aswrltten"has been wr*tln
g”, -------- is displayed (Fig. 7(b)).S
72, only "has written" is displayed, but this can be scrolled by operating the cursor O. The operator selects the marker by operating the cursor keys.
If you move to "en" [Figure 7 (C)] and press the confirmation button,
"Wrote" is changed to "has written". (Figure 7 (d)) Sarayuko In the tense conversion process in Figure 6, when the tense of the translated word determined in 18 is read, the tense of 41 t is changed. K11JriIJ the order of tenses to be displayed according to Tj. Also, the f4 note information other than the tense that the original translated word had is preserved, and the present tense is written as "write" instead of "write".
The current form of s is @has wrltten.”
will be displayed first, then press “write”, @have w
ritten' will be displayed later.

In this way, when a verb is specified, a deformed form of the tense can be automatically generated and replaced, which can significantly reduce the effort required to operate the keyboard for open notes.

Similarly, in the S conversion process, a converted form of the form can be automatically generated using attribute information ψa such as the form of the specified translated word, and can be replaced with 1. FIG. 8 is a diagram showing a screen displaying candidates for converted forms of states.

Similarly, in the state/tense conversion process, attribute information such as the state/tense of the specified translated word is used to automatically generate replacement candidates with simultaneous state and tense conversion.

It can be left behind. FIG. 9 is a diagram showing both sides of candidates for converted forms of cause and tense.

In the same (suunimae near side conversion process 1), an m1 positional replacement candidate that can be replaced is automatically generated using attribute information such as a 11? ring that the specified translation word is a preposition, FIG. 10 is a diagram showing a screen e displaying candidates for replacing prepositions.

Similarly, 1 article conversion process I! I! Here, article replacement candidates that can be replaced are automatically generated using the information that the specified translated word is an article + genus a information such as '11,
(Can run 4 times. Figure 11 shows the crown. IIJ
FIG. 3 is a diagram showing a screen displaying replacement candidates.

Similarly, in singular/plural f depth processing, it is possible to automatically generate and replace possible singular/plural replacement candidates using attribute 1'7V information such as plural change of the designated noun. FIG. 12 shows □□□ showing both sides displaying single/plural replacement candidates.

Furthermore, as shown in FIG.
(1302), the specified translation word (
1301) Automatically determine the part of speech of 1303-130
6), according to each part of speech, perform the process described above and replace the suggestions! ili f display number 1307-1321
) The operator can change A to the confirmed candidate (1322) (1323).

For the translated word selectively specified as described above, the burden on the operator is reduced by automatically generating and replacing 1 place, 1 Okue, and Takuura based on 1 attribute and 1 Seiho. It is possible to reduce qφ and improve the efficiency of translation Kamadokorei.

fj Please note that the present invention is not limited to the embodiments described above. For example, it is possible to automatically generate particle replacement candidates using an English-Japanese machine translation system. In addition to the keyboard, the operator can also use a mouse as a means of operation. Furthermore, the sleeves may not be displayed all at the same time, but may be displayed one by one at the same time as replacement.

In short, the present invention can be implemented with various modifications without departing from its gist.

〔Effect of the invention〕

According to the present invention, by automatically generating ``mori'' (backward candidates) using the attribute information of the translated word, the operator's editing operations can be simplified and the appropriate translated sentence 5: ``to dive'' can be easily obtained. Can be done.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of the present invention. 2nd south is a diagram showing the key arrangement of the input section shown in FIG. 1, FIG. 3 is a diagram showing the screen layout f of the display section shown in FIG. 1, and FIG. 4 is a diagram showing the edit+l1lJ shown in FIG. 1. A flowchart showing the processing flow of the 7513 section, Fig. 5 is similar to Fig. 1
FIG. 6 is a flowchart showing the flow of the processing by the translation unit, FIG. 6 is a flowchart showing the flow of the tense conversion process by the translation unit, and FIG. 7 is a diagram showing the display 1iTii screen of the tense conversion unit 11H. Figure 3 is the transformation process 1! !
Figure 9 shows the display screen for state/tense conversion processing 1, Figure 10 shows the display screen for preposition conversion processing, and Figure 11 shows the display screen for article conversion. The 12th section is a diagram showing the display screen in singular/plural conversion, and FIG. 13 is a flowchart showing the flow of replacement processing. DESCRIPTION OF SYMBOLS 1... Input part k 2... Original text storage part, 3... Translation storage part, 4... Edit control part, 5... Translation part, 6...
・Translation dictionary, 7...Display control section% 8...Display section,
9...Printing department. Agent Patent Attorney Rules Near It Introvert
Matsuyama Hikaru 2Figure 3Figure 2Figure 4Figure 5Figure AFigure 10Figure 11Figure 12

Claims (4)

[Claims] (1) An input section for inputting the original text to be translated;
an original text storage unit that stores the original text, a translation dictionary unit that stores knowledge information used in translation processing, a translation unit that executes translation processing of the original text using the contents of the translation dictionary unit, and a translation processing result. a translation storage unit for storing the obtained translation; a display unit for displaying the original text and the translation; a means for selectively instructing a translated word to be corrected in the translation displayed on the display unit; A machine translation system comprising: an editing section that causes the display section to display replacement candidates for the translated word using attribute information of the translated word. (2) an input section for inputting the original text to be translated;
an original text storage unit that stores the original text, a translation dictionary unit that stores knowledge information used in translation processing, a translation unit that executes translation processing of the original text using the contents of the translation dictionary unit, and a translation processing result. a translation storage unit for storing the obtained translation; a display unit for displaying the original text and the translation; a means for selectively instructing a translated word to be corrected in the translation displayed on the display unit; an editing section that displays replacement candidates for the translated word on the display section using attribute information of the translated word, and the editing section selects the translated word when a replacement candidate is selectively instructed by the instructing means. A machine translation system characterized by replacing with replacement candidates. (3) The machine translation system according to claim 1, wherein morphological change information of the target translation word is used as the attribute information. (4) The machine translation system according to claim 1, wherein part-of-speech information of the translation section to be corrected is used as the attribute information.