JP3324241B2 - Machine translation equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine translation apparatus for efficiently correcting an inappropriate translation in a translation obtained by machine translation.

[0002]

2. Description of the Related Art First, the problem of translation in machine translation, which is the background of the present invention, will be described. English is polysemous,
In particular, when performing English-Japanese translation, translation in word units is one of the major problems in machine translation. For example, sp
The word ring has a completely different concept, such as "spring, spring, spring", but it is very difficult to automatically select an appropriate translation at the current level of machine translation technology.

[0003] In response to English We need the spring.
・ We need a spring.

[0004] We need a spring.・ We need spring. , But this difference is very significant for the user. Humans select appropriate translations while taking into account the overall meaning of the sentence and the positioning of the sentence in the context, but machine translation is a translation on a sentence basis. For the time being, the problem was not easily solved.

Therefore, the user needs to at least correct the translated word with respect to the result of the machine translation. However, the means for correcting the translated word provided by the current translator is not always satisfactory to the user. Absent.

Conventionally, a method of correcting a translated word in a translation result obtained by a machine translation device includes the following. Modify by editor (sequential, batch) (2). Re-translate the entire sentence after registering the user dictionary (3). After translating a single sentence, a post-editing function is used to select the translated word, learn it, and then retranslate the entire sentence. Among the corrections by the editor of (1), the method of correcting the translated word at any time while reading the translated text on the editor is the least efficient. In addition, there is a method based on the function of the editor (search and replace), but the sequential method requires a search and confirmation for each word unit, which takes time. Furthermore, in the case of batch replacement, unnecessary replacement occurs as described below because of a search based on a character string.

[0007] One of the unnecessary replacements is "At that
"At that time, the interval is ~" for English sentences such as "time, the interval is ~"
If the translation of a translation result document containing a translation such as "Time → Time" is requested, even if the request is based on an English word other than time, For replacement,
"The time interval is ~."

[0008] Another unnecessary replacement relies on the translator's incorrect translation selection. For example, suppose the following translation result is in the same document.

The work is going on according to plan. "The work is going as planned." , Each "progress"
The translation "goes forward" is given. Although it is appropriate for the user to say "go" for go,
If the translation of “progress” for “process” should have been “process”, if you perform a face-based batch conversion,
The former states, "The process executes a system call."
The result is reasonable, but as a side effect, the latter is incorrectly translated as "work is being processed as planned." The problem of the batch replacement is that there is no timing at which the user notices the erroneous replacement.

The method of (2) "translating the entire sentence after registering the user dictionary" is often troublesome in registering the user dictionary itself, and has a time cost of retranslating the entire sentence. Too large, (3) "to select a translation by the post-editing function when translating one sentence, to learn and then to re-translate the entire sentence" is that translating one sentence itself is complicated, and that the entire sentence is complicated. The point of retranslating
Same as (2).

[0011]

As described in the "Description of the Related Art", it is not easy to change the translation of a translation result by machine translation, and retranslation takes a lot of time. Further, the user dictionary registration has a problem that the operation is troublesome.

[0012]

Means to Solve the Problems The present invention has been entered
A machine translator for translating an original sentence into a translated sentence, comprising:
Translated sentence storage means for storing sentences, sentence numbers, original words, translated words
A translation information table for storing as a pair of a specified hand <br/> stage for designating a new translation to it and the original word to be changed, in the translation information table the original words specified by the specifying unit as the key and retrieval means for retrieving the original word, a result of the search by the search means, given the original word translated information te
If it is found in the table, the word number containing the found word
The translation of the issue is searched for from the translation storage, and the
Editing means for replacing the translated word with the specified translated word is provided.

[0013]

[Action] In order to store the original words for the translations constituting the translation and to replace the translations based on the information,
The translation is performed appropriately and efficiently without side effects such as search and replacement by a text editor.

[0014]

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

FIG. 1 is a translation information table for storing and managing, in sentence units, what original words are derived from word translations in word units. In the figure, sentence numbers 10 and 20
However, in practice, all sentences have similar translated word information. Each record (line) is composed of three fields (columns), and each field stores a sentence number, an original word (an English word from which a translation is based), and a translated word. That is, each record is
It represents information that "the word XX in the XX line has been translated as XX". Therefore, in the figure, in the sentence 10, "process" is translated as "progress". In the sentence 10, "execute" is translated as "execute". In the sentence 10, the system call is translated as "system call". Yes-In sentence 20, work is translated as "work".-In sentence 20, go is translated as "going forward."-In sentence 20, the plan is translated as "plan." This translation information table is referred to when a certain English word is replaced with another translated word.

FIG. 2 is a functional block diagram of the machine translation apparatus according to one embodiment of the present invention. Reference numeral 1 denotes an original sentence storage unit that stores, as text, an English sentence to be translated, which is taken into the system by a keyboard or the like, and 2 denotes an original sentence storage unit
The translation means 3 for translating the English sentence stored in the translation means 3 translates the result of translation by the translation means 2 as text, and one translation is one record format (a translation for one English sentence is one record. The length is variable for each record, and here the end code of one record is
Assume that there is a line feed code. The translated sentence storage means 4 stores the translated text stored in the translated sentence storage means 3 (in the designated Nth sentence text,
Searches for the specified substring and replaces it with the specified substring. The editing means 5) is a user interface necessary for performing an original text input, a translation instruction, an edit instruction operation, and the like. Reference numeral 6 denotes a translation information storage unit that stores translation information generated in the process of translation by the translation unit 2. The editing unit 4 corrects the translated text in the translated text storage unit 3 while referring to this information.

Next, the procedure (algorithm) of this embodiment will be described with reference to FIG. In step 1A, an original word and a new translated word whose translation is to be changed are obtained. First, in the translation change phase, two data are obtained from the user via the user interface 5; the original word whose translation is desired to be changed; and the translation (new translation) newly desired for the above-mentioned original word. In the above example, the original word for which you want to change the translation is process, and the wrong translation (old translation)
Is a “progress”, and a new desired translation (new translation) is a “process”.

In step 1B, the original word in the translation information table of FIG. 1 is searched. A record in which the value (character string) in the original word column of the translation information table matches the original word character string (process in this example) specified by the user is obtained. From the value in the sentence number column of the obtained record, the sentence number containing the original word can be found. In the example, as a result, it can be seen that the tenth sentence includes process.

Step 1C is a search and replacement of a translated character string in the translated sentence. The tenth translated sentence “progress executes a system call.” Which is the tenth record of the translated sentence storage means 3 is accessed, and the editing means 4 obtains the translated word character string “progress” contained therein. , The character string is replaced with a new translation word “process” obtained from the user. In this embodiment, as a result, “the process executes a system call” is obtained as a new translation.

In the present embodiment, although the 20th sentence includes a translated word “progress” of “the work is proceeding as planned”, the original sentence for the translated sentence is not included in the original sentence. Because of the logic that considers words, this translated sentence is not subject to editing, and erroneous substitution (side effect) seen in a simple character string-based search / replacement by an editor can be suppressed.

Next, another embodiment of the machine translation apparatus of the present invention will be described. FIG. 3 is a functional block diagram of a machine translation apparatus according to another embodiment of the present invention. The roles of the “original text storage means” 1, the “translation means” 2, and the “user interface” 5 are the same as in the above-described embodiment. Reference numeral 7 denotes a translation result storage unit, which has, as described later, a translation of the translation result and translation information of each word constituting the translation as text data. Reference numeral 8 denotes a translated sentence character string generation unit that generates only a translated sentence character string from the translation result storage unit 7. Reference numeral 9 denotes a translation result correction unit that updates the information on the translated word in the translation result storage unit 7.

The method of storing the translation information (data expression) is to write the corresponding original word in word units in the translated sentence character string. Specifically, it is as follows.

The original text to be translated is "You like a ca
"r." is translated into the translation "You like cars." In the conventional method and the above-described embodiment, the original text and the translated text are logically separated from each other by, for example, a file, or by, for example, an odd-numbered line and an even-numbered line. The translation data is clearly maintained by another expression means (table) or the like.

In this embodiment, bilingual information in word units is stored in the translation result (translated text) in a format such as "{you [you] like {cars"}. " Here, symbols such as ｛, ｛, and [,] are not particularly limited thereto. Of course, a sign is used that can be confident that it will not appear in the translation.

Hereinafter, a translation word correction procedure for this data will be described with reference to the flowchart of FIG.

It is assumed that the original word specified by the user and its new translation are "car" and "car", respectively. Therefore, the character string in [] in the translated text is compared with car. If a matching character string is found in [], the character string immediately before [] ({XXX [car]}
XXX part) is replaced with "car".

First, in step 2A, one sentence is extracted from the translation result storage means 9. Suppose that one sentence was "{I [I]} likes {airplane}."

In step 2B, a search is made for the original word car. Since it is not applicable, check for the remaining translation results. Assume that there are two remaining sentences. Therefore, in step 2A, one sentence is extracted from the translation result storage means 9 again.

Next, "{you [you]} is a car [ca
r] I like. Again, in step 2B, a search is made for the original word car. This time, you will find {car}. Therefore, in step 2C, "car" is replaced with "car".

As a result, "{you]
r] I like. Is obtained. Again, step 2B
In the translation result of the same sentence, the original word car is searched. Since the translation result of this sentence no longer includes car as an original word, the translation result of the next sentence is extracted. Of course, at this stage, if there is no remaining translation result, the search / replacement process ends.

Finally, in step 2A,
One sentence is extracted from the translation result storage means 9. next,
"He is a black car.
I like white and white cars. "
And First, "black car" is replaced by "black car" by steps 2B and 2C. Again, in step 2B, {car [car]} corresponding to "white car" is found, and this part is replaced with {car [car]} in step 2C. As a result, "{he [he]} is {black}}
I like {car} and {white} {car}. Is updated to the translation result. The translated sentence character string generating means 8 obtains a translated text “He likes black cars and white cars” by removing the symbols ｛,｝, and [] from the translation result.

[0032]

According to the present invention, an inappropriate translation in a translation obtained as a result of machine translation can be efficiently corrected.

[Brief description of the drawings]

FIG. 1 is a schematic diagram of a translation information table according to an embodiment of the present invention.

FIG. 2 is a functional block diagram of a machine translation apparatus according to an embodiment of the present invention.

FIG. 3 is a functional block diagram of a machine translation apparatus according to another embodiment of the present invention.

FIG. 4 is a flowchart illustrating an implementation procedure according to an embodiment of the present invention.

FIG. 5 is a flowchart illustrating an implementation procedure according to another embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Original sentence storage means 2 Translation means 3 Translation sentence storage means 4 Editing means 5 User interface 6 Translation information 7 Translation result storage means 8 Translated sentence character string generation means 9 Translation result correction means

Claims (2)

(57) [Claims] 1. A machine translation apparatus for translating an input original sentence into a translated sentence, comprising: a translated sentence storage means for storing the translated sentence; and a translation information table for storing sentence numbers, original words, and translated words as pairs. Specifying means for specifying an original word to be changed and a new translation for the same; a searching means for searching for the original word in the translation information table using the original word specified by the specifying means as a key; and the searching means When the specified original word is found in the translation information table as a result of the search, the translated word having the sentence number including the found word is searched from the translated word storage means, and the translated word in the translated word is designated by the designated translated word. A machine translation device comprising editing means for replacing with a translation. 2. A machine translation apparatus for translating an input original sentence into a translated sentence, comprising: a translated sentence storing means for storing a translated sentence and storing an original word associated with a translated word in the translated sentence; A designation means for designating a target original word and a new translation for the target word; a search means for searching for an original word associated with the translation in the translation word storage means using the original word designated by the designation means as a key; A machine translation device including an editing unit that, when a specified original word is found in the translated word storage unit as a result of the search by the searching unit, replaces a translated word associated with the found word with the specified translated word;