JPH0158545B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides an interlingual translation device, in particular, a multilingual translation device configured to have a function of mutually converting between sentences in each language and expressions in an intermediate language. A variable part order code associated with a plurality of variable parts is created in the intermediate language sentence ID in the corresponding memory, and the content of the intermediate language word ID corresponding to the variable part order code can be replaced. The present invention relates to a multilingual translation device that facilitates the translation of sentences with variable parts and enables the translation of fine-grained expressions.

Conventionally, when obtaining a translation device between n types of languages, instead of providing nC _two types of translation functions, a mutual translation function between one language and an intermediate language is provided, and n types of translation functions are provided. It is considered that the above-mentioned translation device is configured according to the following.

Figures 1A and B explain this concept. For example, for Japanese sentence 1 shown in Figure 1A, separate the linguistic sentence part "--, please" and the variable part "kore", and separate them into In the expression of intermediate language 2, the former is associated with intermediate language id "1" and the latter is associated with intermediate language word id "8".
When "this" corresponding to the variable part changes to "that", the variable part "that" is associated with the intermediate language word ID "9", as shown in Figure 1B, and the intermediate language expression " 1, 9”. Of course, from the intermediate language expressions "1, 8" and "1, 9",
It is made to correspond to Japanese sentence 1, English sentence 3, etc.

Although the above concept provides an extremely preferable method, it is not possible to express more detailed content simply by saying, for example, "I will depart from Narita"; instead, it is difficult to express more detailed content, and multiple variables such as "I will depart from Narita at 10 o'clock", etc. It is desirable to be able to translate sentences with parts.
When trying to translate such a sentence, the problem is how to express the existence of multiple variable parts in the intermediate language. Also, for example, "I'll leave Narita at 10o'clock ",
The order in which variable parts occur may change, and it is necessary to take measures to prevent this. Furthermore, it is desirable that the mutual translation mechanisms between each language and the intermediate language representation have the same hardware configuration as much as possible.

The present invention aims to solve the above-mentioned problems, and the multilingual translation device of the present invention converts sentences in each of a plurality of languages into a common intermediate language in correspondence with each of the above languages. In a multilingual translation device that converts into expressions and synthesizes sentences in other languages based on the expressions in the intermediate language, for each of the above languages, a language sentence buffer, and the contents of the language sentence buffer are converted into a language sentence part and a variable part. a string separator/synthesizer that separates and/or synthesizes a language sentence based on the language sentence part and the variable part; A sentence correspondence memory that stores intermediate language sentence IDs in association with each other, and associates variable part word examples of the language with intermediate language word IDs assigned to the intermediate language words corresponding to the word examples. a memory corresponding to variable parts, which stores the intermediate language buffer; the contents of the intermediate language buffer are separated into intermediate language sentence IDs and intermediate language word IDs, and/or the intermediate language sentence IDs and intermediate language word IDs are separated; an intermediate expression synthesis/separation unit that synthesizes; a sentence correspondence memory access unit that accesses the sentence correspondence memory from both the content of the language sentence and the intermediate language sentence ID and reads the contents of the sentence correspondence memory; the variable mentioned above; A variable part corresponding memory that accesses the part corresponding memory from both the variable part and the intermediate language word ID and reads the contents of the variable part corresponding memory.
In addition to providing an access part, a variable part order code that is associated with a plurality of variable parts in the intermediate language sentence ID on the sentence correspondence memory is also provided, and between the first language and the second language. The present invention is characterized in that it has an intermediate language word order replacement processing function that replaces the contents of intermediate language word IDs corresponding to each variable part order code when translating . This will be explained below with reference to the drawings.

FIG. 2 shows the structure of an embodiment of the Japanese sentence to intermediate language expression translation mechanism according to the present invention, and FIG. 3 similarly shows the structure of an embodiment of the English sentence to intermediate language expression translation mechanism.

In FIG. 2, 4 is a translation mechanism section, 5 is a language sentence buffer, and 6 is a character string separation/synthesis section, which separates the contents of the language sentence buffer 5 into a language sentence part and a variable part, or vice versa. 7 is a linguistic sentence, 8 is a first variable part word example, 9 is a second variable part word example, 10 is a sentence correspondence memory, 11 is a variable part correspondence memory, 12 is an intermediate language buffer, and 13 is an intermediate expression synthesis/separation unit which separates the contents of the intermediate language buffer into intermediate language sentence IDs and intermediate language word IDs, or vice versa. 14 is a sentence buffer that searches the sentence correspondence memory 10 based on its contents, or vice versa, sets sentence examples read from the memory 10, and 15 is a sentence buffer. The intermediate language sentence ID read from the memory 10 is set by searching the sentence correspondence memory 10 based on its content, or vice versa, and 16 is a variable part buffer which is variable based on its content. A variable part word example read from the memory 11 is set by searching the part correspondence memory 11, or vice versa.
Reference numeral 17 is a variable part ID buffer in which the intermediate language word ID read from the memory 11 is set based on the contents of the memory for searching the variable part corresponding memory 11, and 18 is a buffer for changing the order of intermediate language words. It represents the processing function section.

In the case shown, in order to facilitate the translation of sentences with multiple variable parts, one example sentence, for example ``-ni-
For example, an ID number such as ``90'' is given to the sentence ``I will depart from ``, and it also indicates that there are multiple variable parts.

Furthermore, on the sentence correspondence memory 10, there is an example sentence "I will depart from - to" and an intermediate language "90,," in the case shown in the figure, taking into account the order of occurrence of variable parts from the perspective of the reference language described later. Store it in pairs with the statement ID.

On the variable part correspondence memory 11, there is an intermediate language word ID "80" corresponding to each variable part word example, for example "10 o'clock".
, and the two are stored as a pair. Furthermore, in the case shown in the figure, an intermediate language buffer 12 is used to facilitate translation of sentences with multiple variable parts.
In the intermediate language expression set above, for example, as shown in the illustration "90, 80, 103", variable part order codes are added corresponding to each intermediate language word ID "80" and "103". Leave it there.

In FIG. 2, when translating the contents of the language sentence buffer 5 into an intermediate language expression, processing is performed as follows.

(1) The contents of the linguistic sentence buffer 5 are separated into a linguistic sentence 7, a first variable part word example 8, and a second variable part word example 9 in the character string separation/synthesis unit 6.

(2) The language sentence 7 is set in the sentence buffer 14, and the sentence correspondence memory 10 is searched by the sentence buffer 14, and in this case, the intermediate language sentence ID "90,..."
is read out to the statement ID buffer 15.

(3) In addition, the first variable part word example 8 is set in the variable part buffer 16, and the variable part correspondence memory 11 is searched by the variable part buffer 16. In this case, the intermediate language word ID "80" is set to the variable part buffer 16. It is set in the part ID buffer.

(4) Similarly, the intermediate language word ID "103" is set in the variable part ID buffer based on the second variable part word example 9.

(5) Each of these IDs is led to the intermediate representation synthesis/separation unit 13, and is associated with the variable part order code in the intermediate language sentence ID "90,," and is processed into the variable part.
The IDs are arranged and set in the intermediate language buffer 12 as "90, 80, 103".

As explained above, the contents of the language sentence buffer 5 are converted into the contents of the intermediate language buffer 12. When conversion into an English sentence is performed based on the contents of the intermediate language buffer 12, it is performed as follows. The order in which the variable parts occur differs between Japanese sentences and English sentences. Taking this into consideration, the intermediate language takes any hypothetical language as the reference language or any one language as the reference language, and based on the order of occurrence of the variable parts in the reference language, the above intermediate language is In the example, "90, 80, 103" corresponds to the contents of the illustrated language sentence buffer 5 in FIG. 2 above. In the case of the Japanese language shown in FIG. 2, it is assumed that the order of occurrence of the variable parts in the reference language is the same as that in Japanese. In the case of the above English sentence, the order of occurrence of the variable parts is different from the order of occurrence of the variable parts in the above reference language. For this purpose, the sentence correspondence memory 10 shown in FIG. 3 stores a form in which the above order is changed, ie, "90, . . ." as the intermediate language sentence ID. Of course, even if there are three or more variable parts, from the point of view of the standard language above, "90,
,,'', and on the other hand, if the order of occurrence of variable parts in English is , then the sentence correspondence memory 10 shown in FIG.
It is sufficient to store it as .

(6) The intermediate language buffer 12 shown in FIG. Intermediate language expression "90, 80 , 103 ” is set.

(7) The intermediate representation synthesis/separation unit 13 uses the intermediate language sentence ID
``90,...'' and intermediate language word IDs ``80'' and ``103'' are separated and set to buffers 15 and 17, respectively.

(8) Based on the contents of each buffer 15 and 17.
The sentence correspondence memory 10 and the variable part correspondence memory 11
is searched, and language sentence example 7 is immediately displayed on sentence buffer 14.
I'll leave at ” is set and the variable part
Variable word examples 8 and 9 on buffer 16
“10o′clock” and “Narita” are set.
Ru.

(9) The character string separation/synthesis unit 6 receives each sentence example and word example and converts it into the sentence “I’ll leave Narita at
10o'clock' is generated and set in language sentence buffer 5.

In the above explanation, based on the order of occurrence of variable parts in the reference language, the order is set as "90,," as the intermediate language word ID, as in the sentence correspondence memory 10 explained with reference to FIG. The replaced items were stored. This is because the information written into the sentence correspondence memory 10 is devised so that the hardware configuration of the translation mechanism section 4 can be made simple and simple. However, if necessary, a conversion memory for changing the order of the variable parts may be prepared to convert the order to suit each language, and the present invention naturally also allows for such an arrangement. It can be considered that it is included in the technical scope.

Although the translation operation from a Japanese sentence to an English sentence has been described with reference to FIGS. It goes without saying that you can freely translate an English sentence into a Japanese sentence by doing this, and furthermore, you can also translate between any one language and any other language by creating an intermediate language expression. Since it is used, it can be done arbitrarily.

It goes without saying that for the separation process in the character string separation/synthesis section 6, grammar dictionaries corresponding to each language are prepared and used.

As explained above, according to the present invention, sentences with multiple variable parts can be freely translated, and detailed expressions can be freely used. Further, the translation mechanism units corresponding to each language can have the same hardware configuration.

[Brief explanation of drawings]

1A and 1B are explanatory diagrams explaining the concept of translation between multiple languages, FIG.
The figure also shows the configuration of an embodiment of the English sentence to intermediate language expression translation mechanism. In the figure, 4 is the translation mechanism section, 5 is the language buffer,
6 is a character string separation/synthesis unit, 10 is a sentence correspondence memory,
11 is variable part compatible memory, 13 is intermediate representation synthesis/
Separation unit, 12 is intermediate language buffer, 14, 15,
16 and 17 represent buffers, respectively.

Claims (1)

[Claims] 1. For each of a plurality of languages, sentences in each language are converted into expressions in a common intermediate language in correspondence with the above languages, and sentences in other languages are converted based on the expressions in the intermediate language. A multilingual translation device that synthesizes a language sentence buffer for each of the above languages, separates the contents of the language sentence buffer into a language sentence part and a variable part, and/or converts the language into a language based on the language sentence part and the variable part. a string separation/synthesis unit that synthesizes sentences; a sentence correspondence memory that stores sentence examples in the language in association with intermediate language sentence IDs assigned to sentences in the intermediate language that correspond to the sentence examples; A memory corresponding to a variable part that stores a variable part word example of a language in association with an intermediate language word ID assigned to a word of the intermediate language corresponding to the word example, the intermediate language buffer, and the intermediate language. an intermediate expression synthesis/separation unit that separates the content of the buffer into an intermediate language sentence ID and an intermediate language word ID and/or synthesizes the intermediate language sentence ID and intermediate language word ID; A sentence corresponding memory access unit that accesses the contents of the memory corresponding to the sentence by accessing both the content and the intermediate language sentence ID, and reading the contents of the memory corresponding to the sentence from both the variable part and the intermediate language word ID. Variable part memory that accesses and reads the contents of the variable part memory.
In addition to providing an access part, a variable part order code that is associated with a plurality of variable parts in the intermediate language sentence ID on the sentence correspondence memory is also provided, and between the first language and the second language. 1. A multilingual interlingual translation device characterized by having an intermediate language word order swapping function for swapping the contents of intermediate language word IDs corresponding to each of the variable part order codes when translating.