JP2000123012A - Machine translation device and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reflect the non-language information of the word of first language on second language. SOLUTION: This device is provided with a non-language information reflections setting part 8 for preliminarily setting non-language information for reference and a translation processing part 9. This translation processing part 9 is provided with a function for reflecting the non-language information on the word of second language being the translated result of first language to which the non-language information is added according to the length of character modification, phoneme range, or number of characters of the word of the first language at the time of fetching the first language from a transmission line 1.

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 and a recording medium for reflecting non-verbal information attached to a first language in a second language.

[0002]

2. Description of the Related Art In recent years, using a WWW browser, an HTM
2. Description of the Related Art Machine translators that reflect non-verbal information such as layouts in translation results, such as WWW translators that browse and translate L-form Internet information, are on the market. Also,
The same applies to general English-Japanese translators, and there are many requests to reflect non-linguistic information added to a first language or the like in a translation result.

In such a machine translator, the word "School" in the first language can recognize the word "school" in the second language from the translation dictionary and the word "School" in the first language can be recognized. If the entire "School" is italicized, it can be translated into the second language word "school" and the translation result "school"
Can be reflected in italics and the like.

[0004]

However, in a general English-Japanese translator including a WWW translator, if only "Sch" in "School" is in italics, for example, the translation result of Japanese in Japanese is used. Italics cannot be reflected in parts. That is, when one word in the first language includes a modifier character based on non-linguistic information in a range smaller than the word, the non-linguistic information cannot be reflected in the second language.

The present invention has been made in view of the above circumstances, and has as its object to provide a machine translation apparatus that appropriately reflects non-linguistic information attached to a first language in a translation result. It is another object of the present invention to provide a recording medium having a translation processing program for appropriately reflecting non-linguistic information attached to a first language in a translation result.

[0006]

In order to solve the above-mentioned problems, the present invention provides a non-verbal information reflection setting means for setting non-verbal information for reference in advance, and a method in which non-verbal information is included in the word of the first language. It is determined from the non-linguistic information of the non-linguistic information reflection setting means whether or not it is added, and when the non-linguistic information is added, the non-linguistic information is set according to the position of the modifier character of the word in the first language. A translation processing means for reflecting the non-linguistic information on a word in a second language which is a translation result of one language, or a non-linguistic information reflection setting means for setting reference non-linguistic information in advance; It is determined from the non-linguistic information of the non-linguistic information reflection setting means whether or not non-linguistic information is added to the word of one language, and when the non-linguistic information is added, the word of the first language is determined. In words of first language by length and non-verbal information A machine translation apparatus provided with translation processing means for reflecting the non-linguistic information in a word of a second language, which is a translation result of the first language, in accordance with a ratio to a length of a modifier character and a position of the modifier character It is.

In another aspect of the present invention, it is determined whether or not non-linguistic information is added to a word in a first language from non-linguistic information of a non-linguistic information reflection setting unit, and the non-linguistic information is added. A configuration provided with translation processing means for reflecting the non-linguistic information in a word in a second language, which is a translation result in the first language, according to a range of phonemes of a modifier character in the word, or When the non-linguistic information is added to the word in the first language, the translation result of the first language is a second translation result of the first language according to the number of modified characters and the modified character position with respect to the total number of characters in the word by the non-linguistic information A translation processing means for reflecting the non-linguistic information on the words of the language is provided.

[0008] Therefore, the present invention, by taking the above means, even if a modified character by non-linguistic information is described in a part of the word of the first language, the ratio and position of the modified character , Or according to the range of modifier phonemes, and according to the number and position of modifier characters in the word, non-verbal information can be reflected in words in the second language, and the intent of the first language It is possible to reflect exactly what you do in the second language.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. (Embodiment 1) FIG. 1 is a block diagram illustrating an embodiment of a machine translation apparatus and a recording medium according to the present invention.

[0010] This machine translation apparatus has a data transmission line 1
Is connected to the data transmission line 1 here.
It does not need to be the Internet, but refers to an intranet or other various data transmission lines. Further, a case where the translated original is read by using a scanner or a case where the translated original is received via a communication satellite may be used. In this case, connection to the data transmission line 1 is not necessarily required.

The machine translation device is connected to the data transmission line 1 via the communication control unit 2 when fetching a translation original from the data transmission line 1. This communication control unit 2
Different protocols "http", "smtp", and "ML" are used in accordance with ML format documents, other markup documents, e-mail documents, and file transfer data.
Based on “FTP”, the translation original on the data transmission line 1 is taken in.

This machine translation apparatus includes a communication control unit 2 (or a receiving unit corresponding to a scanner or a communication satellite), and a pointing device such as a keyboard or a mouse for performing an instruction to take in a translation original, an instruction to translate, and other necessary operations. Input device 3, a display device 4 such as a CRT for displaying captured original text information and translation results, a bit map, a wall projection, a translation processing recording medium 5 for recording a translation processing program, and storing dictionary / grammar data. A dictionary / grammar storage unit 6, a document data storage unit 7 for storing the information of the fetched translation original and the translation result, and a reference non-language in advance for determining whether the information attached to the word of the translation original is non-linguistic information. A non-language information reflection setting unit 8 for setting information and a translation processing program recorded on the translation processing recording medium 5 are read and a predetermined translation process is performed. Run, is constituted by constituted translation processing unit 9 by the CPU.

As the recording medium 5, a magnetic disk, a magnetic tape, and a CD-ROM are used. In addition, a DVD-ROM, a floppy disk, an MO,
D, CD-R, memory card, etc. may be used.

The dictionary / grammar storage unit 6, document data storage unit 7, and non-language information reflection setting unit 8 may be stored in the same recording medium by dividing the area. As the reference non-language information set in the non-language information reflection setting unit 8, for example, the following tags are set in the case of an HTML format document.

For example, italic: <I> ××× </ I> Bold: <B> ××× </ B> Text size: <FONT SIZE = n> ×× <FONT SIZE = n ′> Color: <FONT COLOR = n> ××× <FONT COLOR = n ′> Others, erased characters, monospace, underline, superscript, subscript,
Tags such as emphasis are set.

In the case of an electronic mail document, for example, a control code or the like corresponding to each piece of non-language information is set. The point is that if some of the modifiers of each word in the document are recognized by non-verbal information, HT
Not limited to ML format documents and e-mail documents.

Further, the translation processing unit 9 includes H
W to read Internet information in TML format
It has software and components for taking in documents as needed, such as the need for a WW browser.

The translation processing unit 9 reads the original translation, which is the first language, from the transmission line 1 or a scanner or the like, and determines whether or not non-linguistic information is added to each word of the original translation. Language information presence / absence determination function 11,
When the non-language information presence / absence determination function 11 determines that the word has non-language information, the non-language information is deleted from the word, the non-language information content and the character modification position are specified, and the document data recording unit 7 A translation processing function 13 for translating the words of the translation original from which the non-language information has been deleted as described above into a second language; and a translation processing function 13 for translating the words of the translation original according to the non-language information. Determines whether the character modification range is more than a word or less than a word, and if it is within a word, calculates the character modification ratio from the character modification length of the word and the length of the word from which non-linguistic information has been deleted Means 14 and a non-language reflection device for reflecting non-language information in the second language translated by the translation processing function 13 in accordance with the ratio and the modified character position calculated by the modification ratio calculation means 14 It is constituted by 15.

Next, the operation of the above apparatus will be described with reference to FIG. The translation processing unit 9 reads the translation processing program from the translation processing recording medium 5 and implements the following functions. That is, the translation original (first language) A described in, for example, the HTML format is read from the transmission line 1 via the communication control unit 2 (S1), stored in the document data storage unit 7, and stored in the memory counter. = 1 is set (S2), the i-th word Ai of the translation original A is read (S3), and the non-language information is added to the word Ai with reference to the non-language information set in the non-language information reflection setting unit 8. It is determined whether or not it has been performed (S4). These steps S1 to
S4 implements the non-language information presence / absence determination function 12. As a result of this determination, when non-verbal information is not added to the word,
A normal translation process is performed (S5).

By the way, it is assumed that the following non-linguistic information is added to the i-th word Ai of the translated original sentence A described in the HTML format. In this case, it is determined that non-verbal information is added to the word in step S5.

<BODY><I> math </ I> ematician <P><BODY> This word means that "math" is italicized from the italic tag <I> math </ I> . Therefore, when it is determined that the word is non-linguistic information-added word, the non-linguistic information <I>, </ I> is deleted from the word, mathematician = Aj, and the non-linguistic information contents <I>, </ I> and the place of the modifier character "first to fourth" are Bi,
It is stored in the document data recording unit 7 (S6, non-language specific data storage function 12).

Then, the word Aj from which the non-linguistic information has been deleted
Is translated into a second language, for example, "Mathematician" in Japanese using the dictionary and grammar of the dictionary / grammar storage unit 6, and the translation result is set to Cj (S5, S7, translation processing function).

Next, it is determined whether or not the character modification range of the word based on the non-verbal information is equal to or larger than the word from Bi (S8). If the character modification range is equal to or larger than the word, all the "mathematicians" which are translation results Cj are deleted. Non-linguistic information is reflected in italics based on <I> and </ I> of the linguistic information contents (S9).

On the other hand, when the character modification range of a word based on non-verbal information is within a word, the modification start position Bb, the modification end position Be and the modification end position "math" (italic) of the character modification in Ai are based on Bi. The modification length Dl is determined (S1
0), and the length of Aj is determined (S11). Then, the ratio of Dl is calculated from Dl and the length of Aj (S12). These steps S10 to S12 implement the modification ratio calculating means 13.

Once the ratio is obtained, the first character "number", which is a part of "mathematician", is expressed in italics in accordance with the ratio, for example, 1/3 and Bi. Language information is reflected in the translation result (S13,
Non-language reflection function).

Thereafter, it is determined whether or not there is a next original sentence word (S14). If there is an original sentence word, in step S15, i = i + 1 is incremented in step S3.
And the same processing is repeated.

By the way, among the "mathematician", in the case of a modifier character in which the first four characters "math" are italic, the length of this "math" and the length of "mathematician" are given by:
When "math" is 31%, italic is reflected in the characters of the translation result of 31% from the beginning. In other words, only the “number” corresponding to 31% from the beginning of the “mathematician”, which is the translation result, is made italic, so that the non-verbal information attached to the original word is reflected in the translation result.

Therefore, according to the above embodiment,
If there is non-linguistic information in a word of the first language, after recording the semantic content of the non-linguistic information and the location of the character modification range, the non-linguistic information is deleted, and only words in the original language are translated. After that, the character modification ratio is calculated from the length of the character modified by the non-verbal information and the length of the entire word, and the non-verbal information is reflected in the translation result according to the ratio and the character modification position. The intention can be faithfully reflected in the translation result.

This translation device is effective when the character modification range based on non-linguistic information is completed within a word. That is, the position where the modification range occupies in the word is calculated. For example, if the character modification range is 25% to 50% from the beginning, 25% from the beginning of the word included in the translation result.
The original non-verbal information is reflected on characters in the range of% to 50%. (Embodiment 2) FIG. 3 is a block diagram illustrating a machine translation apparatus and a recording medium according to an embodiment of the present invention.

This machine translation device is connected to a data transmission line 1 via, for example, a communication control unit 2 as in FIG.
An input device 3 such as a keyboard and a pointing device such as a mouse for performing an instruction for taking in the original information, a translation instruction and other necessary operations, and a display device 4 such as a CRT for displaying the acquired original information and a translation result, a bit map, and a wall projection. A translation processing recording medium 21 for storing a translation processing program, a dictionary / grammar storage section 6 for storing dictionary / grammar data, a document data storage section 7 for storing the acquired original text information and information on translation results, The non-language information reflection setting unit 8 that sets non-language information for reference in advance so that it is possible to determine whether or not the information attached to the original text information is non-linguistic information is recorded in the translation processing recording medium 5. Reads a translation processing program and executes a predetermined translation processing, C
It is composed of a translation processing unit 22 composed of PU.

Note that the type of the recording medium 21, the same recording medium as the dictionary / grammar storage unit 6 and the non-linguistic information reflection setting unit 8, and the like are the same as in FIG. 1 and will not be described here. The translation processing unit 22 fetches a translated original sentence in the first language from the transmission line 1 or a scanner or the like, and determines whether or not non-linguistic information is added to each word of the translated original sentence. When the function 31 and the non-language information presence / absence determination function 31 determine that the word has non-language information,
A non-language specifying data storage function 32 for deleting the non-language information from the word, specifying the non-language information content and its character modification location, and storing the same in the document data recording unit 7; A translation processing function 33 for translating into two languages, and furthermore, when the character modification range is within a word,
Phoneme decomposition processing function 3 for decomposing the word into phonemes and determining the number of phonemes in the character modification range based on the non-verbal information
4 and a non-language reflection function 35 for character-modifying the character of the translation result according to the content of the non-linguistic information according to the number of the phoneme determined by the phoneme decomposition processing function 34.

Next, the operation of the above apparatus will be described with reference to FIG. The translation processing unit 22 reads the translation processing program from the translation processing recording medium 21 and implements the following functions. That is, the translation original (first language) A described in the HTML format, for example, is read from the transmission line 1 via the communication control unit 2 (S21), stored in the document data storage unit 7, and then stored in the memory counter. =
1 is set (S22), and the i-th word Ai of the translated original text A is set.
Is read (S23), and it is determined whether or not non-linguistic information is added to the word Ai by referring to the non-linguistic information set in the non-linguistic information reflection setting section 8 (S24). These steps S21 to S24 implement the non-language information presence / absence determination function 31. If the result of this determination is that non-linguistic information has not been added to the word, normal translation processing is performed (S25).

By the way, it is assumed that the following non-language information is added to the i-th word Ai of the translated original sentence A described in the HTML format. In this case, it is determined in step S25 that non-verbal information has been added to the word.

<BODY><NOBR> mathe <B> ma </ B> ticion </ NOBR><BODY> Here, non-verbal information <NOBR>, <B>,
</ B> and </ NOBR> are deleted, mathematician = Aj, and the non-verbal information contents <B>, </ B> and the location (position) "6th and 7th from the beginning" are set to Bi. The data is stored in the data recording unit 7 (S26, non-language specific data storage function 32). Thereafter, the word A from which the non-verbal information is deleted
j is translated into a second language, for example, "Mathematician" in Japanese, using the dictionary and grammar of the dictionary / grammar storage unit 6, and the translation result is Cj (S27, translation processing function 33).

Next, it is determined whether or not the modification range of the word Ai is a word or more from the non-verbal information content and the character modification location (position) Bi (S28). .. </ B>, all words of the translation result in the second language are reflected in bold (S29).

However, when the modification range of the word Ai is a partial modification in the word Ai, the modification start position of the character modification "math" (italic) in Ai is set to Bb, based on Bi.
The modification end position is set to Be (S30), and Aj = ma
Regarding thematician, phoneme decomposition such as "math.e.ma.ti.cian" is performed, and it is confirmed that the character "ma" in bold is the third phoneme, which is defined as Gi (S31). In these steps S30 and S31, a phoneme decomposition processing function 34 is realized.

Further, it is determined from the order of the translation result that Gi including the modification start position Bb and the modification end position Be is Gb and Ge (S32). Then, a word Fj corresponding to Dj = "ma" in Cj = "mathematician" as a translation result is extracted (S33), and the Geth character from Gb in this Fj is modified to bold (S34). In these steps S32 to S34, the non-language information reflecting function 35 is realized.

Thereafter, it is determined whether or not there is a next original sentence word (S35). If there is an original sentence word, the process returns to step S23 by repeating i = i + 1 (S36), and the same processing is repeated.

Therefore, according to the above-described embodiment, it is determined whether or not non-linguistic information exists in the translated original text. If the non-linguistic information does not exist, normal translation processing is executed. After storing the non-verbal information content and its location, the non-linguistic information is deleted from the translated original and translated.
At this time, the content of the non-verbal information is reflected in the translation result,
If the non-verbal information extends beyond the word, it is reflected in the whole word. If the non-linguistic information exists in a part of the word, the phoneme is decomposed, and it is determined whether or not the modification range of the word is included in the phoneme. I do. For example, if the range is from the second to the fourth counting from the beginning, the contents of the non-verbal information are reflected in the second to fourth characters from the beginning of the word corresponding to the phrase included in the translation result. Is to be modified as follows.

For example, if there is a sentence “He is a mathematician”, “mathematician” is “math · e · ma
If the phoneme is decomposed as “ti ・ cian” and only the first phoneme is found to have italic modification, the translation result “He is a mathematician.” Only one character becomes the italic translation result. (Other Embodiments) In the first embodiment, the ratio of the character modification range (length) in a word based on non-verbal information is obtained, and the ratio and the character modification position in the word are reflected in the translation result. However, simply by counting the number of modifier characters with respect to the number of characters in the entire original translated text based on the non-linguistic information, the non-linguistic information may be reflected in the character of the translation result from the relationship of the modifier character position.

For example, in the case of the word "Mathematician", only the first character is a large font.
Accordingly, the font is reflected so that only the first character of the "mathematician" as a translation result becomes a large font.

[0042]

As described above, according to the present invention, even if a part of a word in the first language has a modifier character by non-linguistic information, the translation result of the non-linguistic information And the intention of the first language can be faithfully expressed in the second language. Further, by using the translation processing program of the recording medium according to the present invention, the non-linguistic information attached to the first language can be appropriately reflected in the translation result.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing an embodiment of a machine translation device according to the present invention.

FIG. 2 is a flowchart of execution of a translation processing program recorded on a recording medium.

FIG. 3 is a configuration diagram showing another embodiment of the machine translation device according to the present invention.

FIG. 4 is a flowchart of execution of a translation processing program recorded on a recording medium.

[Explanation of symbols]

 5, 21 recording medium 8 non-verbal information reflection setting unit 9, 22 translation processing unit 14 modification ratio calculation function 34 phoneme decomposition processing function

Claims (7)

[Claims] 1. A machine translation apparatus for reflecting non-linguistic information attached to a first language in a second language, a non-linguistic information reflection setting means for setting reference non-linguistic information in advance, the first language It is determined from the non-linguistic information of the non-linguistic information reflection setting means whether or not non-linguistic information is added to the word, and when the non-linguistic information is added, the modifier character of the word in the first language A translation processing means for reflecting the non-linguistic information in words of a second language, which is a translation result of the first language, according to the position of the first language. 2. A machine translation device for reflecting non-linguistic information attached to a first language in a second language, a non-linguistic information reflection setting means for setting reference non-linguistic information in advance, the first language It is determined from the non-linguistic information of the non-linguistic information reflection setting means whether or not non-linguistic information is added to the word, and when the non-linguistic information is added, the length of the word in the first language and the non-linguistic information are determined. The non-verbal information is reflected in the second language word that is the translation result of the first language according to the ratio of the modifier character length and the modifier character position in the word of the first language according to the linguistic information. And a translation processing means. 3. A machine translation device for reflecting non-linguistic information attached to a first language in a second language, wherein the non-linguistic information reflection setting means sets in advance non-linguistic information for reference; It is determined from the non-linguistic information of the non-linguistic information reflection setting means whether or not non-linguistic information is added to the word, and when the non-linguistic information is added, the range of the phoneme of the modified character in the word is determined. And a translation processing unit for reflecting the non-linguistic information in a word of a second language, which is a translation result of the first language. 4. A machine translation apparatus for reflecting non-linguistic information attached to a first language in a second language, a non-linguistic information reflection setting means for setting reference non-linguistic information in advance, the first language It is determined from the non-linguistic information of the non-linguistic information reflection setting means whether or not non-linguistic information is added to the word, and when the non-linguistic information is added, the number of modifier characters to the total number of characters in the word by the non-linguistic information And a translation processing means for reflecting the non-linguistic information in a word of a second language, which is a translation result of the first language, in accordance with the character position of the first language. 5. A recording medium storing a translation processing program for reflecting non-linguistic information attached to a first language in a second language, wherein the computer stores the non-linguistic information in words of the first language. A non-verbal information presence / absence determining function for determining whether or not the word has been added; and a non-verbal information content and character modification position attached to the word when the non-verbal information presence / absence determining function determines that the word has non-verbal information. A non-language specific data storage function for capturing and storing, a translation processing function for translating a word in the first language from which the non-language information has been deleted into a second language, and a modifier character for the word in the first language A computer that stores a translation processing program for realizing a non-language reflection function of reflecting the non-language information on a word in a second language that is a translation result of the first language according to the position of the first language Only take a recording medium that can be. 6. A recording medium storing a translation processing program for reflecting non-linguistic information attached to a first language in a second language, wherein the non-linguistic information is stored in a computer in words of the first language. A non-verbal information presence / absence determining function for determining whether or not the word has been added; and a non-verbal information content and character modification position attached to the word when the non-verbal information presence / absence determining function determines that the word has non-verbal information. A non-language specific data storage function to capture and store, a translation processing function to translate the first language word from which the non-language information is deleted into a second language, and a length of the first language word. A modification ratio calculation function for calculating a ratio of a modification character length in a word of a first language according to non-linguistic information; and a modification ratio and a modification character position according to the modification ratio calculation function, the first language Translation of Second language computer readable recording medium storing a translation processing program in order to realize a non-language reflecting function of the reflecting nonverbal information words of a fruit. 7. A recording medium storing a translation processing program for reflecting non-linguistic information attached to a first language in a second language, wherein the non-linguistic information is stored in a computer in words of the first language. A non-verbal information presence / absence determining function for determining whether or not the word has been added; and a non-verbal information content and character modification position attached to the word when the non-verbal information presence / absence determining function determines that the word has non-verbal information. A non-language specific data storage function for taking in and storing, a translation processing function for translating a word in the first language from which the non-language information has been deleted into a second language, and converting the word in the first language into a phoneme. The phoneme decomposition processing function for decomposing and specifying the phoneme range of the modified character in the word, and the second language translation result of the first language according to the phoneme range specified by the phoneme decomposition processing function. Language Machine translation apparatus is characterized by providing a translation processing means for reflecting the non-verbal information word.