JPH0434644A - Machine translation system - Google Patents

Abstract

PURPOSE:To efficiently execute translation processing by deleting a blank part in an unknown word which can not be retrieved in an original sentence or substitute a code for the blank part and executing word retrieval again. CONSTITUTION:The machine translation system is constituted of an input part 1, an original sentence storing part 2, a translated sentence storing part 3, an edition control part 4, a translation part 5, a translation dictionary 6, a display control part 7, a display part 8, and a printing part 9. The translation part 5 deletes a blank part in an unknown word which can not be retrieved at the time of executing word retrieval for retrieving words corresponding to an original sentence while referring the dictionary 6 or substiture a code for the blank part and then executes word retrieval again. Even when a word including a space exists in an inputted original sentence, the word is surely retrieved and translation processing is efficiently executed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Purpose of the Invention (Industrial Application Field) The present invention relates to a machine translation system that improves word search for searching for headwords in a translation dictionary from words in an input original text.

(Prior Art) In recent years, systems that automatically machine translate input original texts using computers have been attracting attention.

Conventional machine translation systems basically divide the input source text into predetermined processing units such as words (phrases) by looking them up in a dictionary, morphologically analyzing them, and parsing their syntax, and then search a translation dictionary for each processing unit. The system is configured to find corresponding translated words (translated phrases), etc., and combine them according to predetermined translation rules to obtain the translated text. In dictionary lookup processing and morphological processing, input character strings are usually matched with dictionary entries using grammatical transformation information such as pragmatic inflection rules and suffix inflection rules. "talk", "talk"
If it matches a dictionary entry in the translation dictionary, it means that the search was successful. Then, later syntax analysis etc. can be performed. Further, a character string that is not recognized as a word by dictionary lookup or morphological processing will be processed as an unknown word.

However, in actual documents, there are cases where the grammatical processing described above cannot be used.
In some cases, "a' is used as a symbol to represent a blank space". Furthermore, as in the title of a text, a blank space (hereinafter referred to as a space) may be intentionally inserted between the words rW ORL DJ for layout purposes. In addition to this, for example, “Defect
When connecting compound words such as ``skipping,'' some operators may insert spaces, such as ``defect skipping.'' When a sentence containing these words is input into the machine translation system, the words become unknown words even though the words are originally stored in the translation dictionary. Therefore, there was a problem that translation processing could not be performed efficiently.

(Problems to be Solved by the Invention) As described above, in conventional machine translation systems, when matching words in an input source text with dictionary entries in a translation dictionary, it is possible to detect words that include spaces in the source text. However, since word searches cannot be performed, there is a problem in that translation processing cannot be carried out efficiently.

The present invention has been made in order to solve the above-mentioned problems, and is a machine translation system that can reliably perform word searches even if there are words that include spaces in the input original text, and can therefore perform translation processing efficiently. The purpose is to provide a system.

[Structure of the Invention] (Means for Solving the Problems) In order to achieve the above-mentioned purpose, the machine translation system of the present invention includes an input section for inputting an original text to be translated, and knowledge information used for translation processing. a translation dictionary section that stores the translation dictionary section; a translation section that uses the translation dictionary section to translate the original text to obtain a translated text; and a display section that displays the original text from the input section and the translated text from the translation section. . Then, when performing a word search for the original text by referring to the translation dictionary section to search for a corresponding word, the translation section deletes blank spaces in the unknown words or replaces them with symbols for unknown words that cannot be searched. The word search is performed again.

Further, the input section is for instructing whether the translation section should delete the blank section in the unknown word or replace it with a symbol.

(Function) According to the machine translation system of the present invention configured as described above, when the translation unit uses the translation dictionary to search for words in the original text from the input unit, unknown words are detected. If there is a space in the word, the space is deleted or replaced with a symbol.

As a result, the translation dictionary can correctly search for words in the original text without modifying them. Therefore, translation processing can be performed efficiently.

(Example) Hereinafter, an example of the present invention will be described in detail with reference to the drawings.

FIG. 1 is an overall block diagram of a Japanese-English machine translation system as an embodiment of the present invention. The system includes an input section 1 consisting of a keyboard, an original text storage section 2 that stores the human-keyed Japanese text as the original text, a translation storage section 3 that stores the English text corresponding to the input Japanese text as the translated text, and controls the entire system. an editing control section 4 that executes the translation process; a translation section 5 that executes the translation process;
, a translation dictionary 6 containing knowledge information used in translation processing;
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 transcribing the original text, translated text, etc. and outputting it as 7)-tocoby.

FIG. 2 is a diagram showing an example of the key layout of the input unit 1.

The human resources section 1 is equipped with the following various keys in addition to character keys for Japanese and English human resources.

Translation instruction key = 100 Edit key: 101-106 Function key: 110-117 Cursor key: 130-133 Space deletion mode key: 140 Space - midpoint conversion - 8141 Mode kiss base - hyphen conversion, 142 Mode key Figure 3 is 3 is a diagram showing an example of a layout of a screen of a display unit 8. FIG. 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 dialog translation processing by the editing control unit 4. In interactive translation processing, the operator can input the original text and edit the corresponding translated text as appropriate. The editing control unit 4 performs steps 81 to S6.
In this step, 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, a corresponding process is performed.

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 S12.

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 Sll.

When the operator operates the cursor key 130 or other control keys, processing for moving the cursor corresponding to the key operation and other processing are performed in step S13 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, and then the corresponding character code is sequentially set in the input buffer (not shown) in the editing control unit 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). When the operator presses the translation instruction key 100 at any time, for example, when he has finished inputting a - sentence, this key is detected in step S2, and the editing control unit 4 sends the translation unit 5 to the original text in the input buffer. is supplied to instruct the start of translation processing (step 510). In addition, if input editing such as correction, insertion, deletion, etc. is required in the middle of editing the original text, move the cursor to the desired editing location using the cursor keys, and then press the insert key 101 and the delete key 10.
This can be done using the 2nd etc. edit key.

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

The morphological analysis unit (S51) uses the conjugation change/dictionary 6a to convert words with changes in endings, etc. into their original forms (basic forms) in the input original text.

The dictionary search section (S52) searches the word/idiom dictionary 6b for each word constituting the input original text, and collects information such as its part of speech and translated words.

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

The syntax analysis unit (S54) analyzes the syntax of the translation candidate using the analysis grammar 6c, 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,
The process returns to the syntax analysis section (S54).

The syntax generation unit (35B) determines the English word order from the English structure according to the generative grammar 6e and converts it into a word string.

The morpheme generation unit (S57) completes the translated sentence by changing the word form etc. using the morpheme generation grammar 6f.

Each of the above-mentioned processing units outputs a signal to the editing control unit 4 when starting processing.

Referring again to FIG. 4, when the editing control section 4 detects the translation completion signal in step S1, the process proceeds to step S7, where it sends the translation candidate and auxiliary information obtained from the translation section 5 to the display control section 7 and causes the translation to be displayed. (Step S7). The display control unit 7 displays the translation candidate in the translation display area at a position corresponding to the input original text (on the left side of the screen), and if there is auxiliary information, displays the translation with high brightness or in reverse video so that the operator can I'm trying to identify that.

From this state, when the edit key is pressed, the edit control unit 4
detects this in step S4, and performs editing processing corresponding to each key on the word or phrase indicated by the cursor at this time (step 511). For example, insert key 10
Operation 1 inserts a character before the cursor position.

By operating the delete key 102, the character string in the range indicated by the cursor is deleted. By operating the movement key 103, the range indicated by the cursor is moved.

By operating the cancel key 104, the above 101°102.1
Nullifies the effect of 03. Modifications are made by operating the key 105, using auxiliary information to display other modification candidates for the word or phrase indicated by the cursor.

Note that the following functions can be realized using the function keys.

By operating the translated word display key 110, the translated word can be displayed for each word in the translated sentence. By operating the dictionary display key 111, a dictionary can be displayed for a word in the original text. By operating the dictionary registration key 112, words and phrases can be registered. By operating the dictionary deletion key 113, words and phrases registered through dictionary registration can be deleted. By operating the partial translation key 114, when translation fails, the partial translation is displayed.

Further, the cursor keys include cursor movement keys 130 for moving the cursor in each direction, a unit switching key 131 for switching the unit in which the cursor moves, and an inter-area movement key 132 for moving the cursor between display areas.
, and zoom keys to reduce the cursor size character by character or enlarge the cursor size word by word. Then, the space deletion mode key 140 is used for the dictionary lookup,
In morpheme processing, 0N10FF specifies whether word search should also be performed for character strings with spaces removed from the input sentence. The space←→midpoint conversion mode key 141 specifies with 0N10FF whether to perform a word search after replacing spaces in the input sentence with midpoints in the dictionary lookup and morpheme processing. The space□hyphen conversion mode key 142 selects whether to replace spaces in the input sentence with hyphens and then perform a word search in the dictionary, lookup, and morpheme processing.
Specify with N10FF.

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

In the system configured as described above, a word search in a dictionary will be further explained in detail below.

FIG. 6 is a diagram showing the flow of word search processing. S8
In step 1, a word search is performed for each sentence in the original text to see if there is a word that corresponds to the headword in the translation dictionary, and unknown words are extracted. As a result, the matched headword is registered as a word candidate in the dictionary word list, and the word search process ends.

Words that are not set in the dictionary word list in S61 are regarded as unknown words and the next step is performed. Next, in 862, it is determined whether the space deletion mode set by the space deletion mode key 140 is ON or OFF;
If F, go to SB4.

If it is ON, matching is performed in S63 between the character string obtained by removing spaces from the input character string and the headword in the dictionary. As a result, the matched headword is registered as a word candidate in the dictionary word list.

Next, in 564, it is determined whether the space-midpoint conversion mode set by the space-midpoint conversion mode key 141 is ON or OFF, and if it is OFF, the process goes to SB5. If it is ON, matching is performed in S65 between the character string in which spaces in the input character string are replaced with midpoints and the headword in the dictionary. As a result, the matched headword is registered as a word candidate in the dictionary word list. Next, in SBB, the space←→hyphen conversion mode set by the space←→hyphen conversion mode key 142 is turned on.
It is determined whether or not it is OFF, and if it is OFF, the process ends.

If it is ON, matching is performed in step 866 between the character string in which spaces in the input character string are replaced with hyphens and the headword in the dictionary. As a result, the matched headword is registered as a word candidate in the dictionary word list.

The above processing will be explained below using a specific example.

FIG. 7 is an example of entry words and translated words in a translation dictionary.

FIG. 8 is a diagram for explaining the processing when "disfu" is given as an input character string containing unnecessary spaces. Space deletion mode is now ON
Assume that the space←-midpoint conversion mode and the space-hyphen conversion mode are OFF. Note that the input character string "disk" is incorrect. First, at 860 in FIG. 6, the input character string is matched with the dictionary in FIG.
Since there is no entry word that matches ``sufu'' etc.
Nothing is set in the dictionary word list in the figure (8a).

Since nothing is set in the dictionary word list in S61, the word is determined to be an unknown word and the process proceeds to SB2. Next 88
51 because the space deletion mode is ON in 2.
33 will be carried out. Here, the input string "day sufu"
The character string "disc" with spaces removed from it becomes the new dictionary matching target character string. When this character string is matched with the dictionary, "disc" is matched as shown in FIG. 7, and this word is registered in the dictionary word list (8b). Here, the number in parentheses after "disk" is the corresponding dictionary number. The process is S64. S
After passing through 8B, we get "disc" from "disk" as the dictionary lookup result.

FIG. 9 is a diagram for explaining processing for the character string "defect skipping."

Currently, space deletion mode and space-hyphen conversion mode are OFF, and space-midpoint conversion mode is ON.
Suppose that First, at 880 in Figure 6, the input character string is matched with the dictionary in Figure 7, but since the dictionary does not have headwords that match "defect skipping,""defect,""skipping," etc., the dictionary word Nothing is set in the list (9a). Then, in S61, it is determined that it is an unknown word, and in S62, a space deletion conversion is confirmed. This causes the process to proceed to S84. In SB4, since the space-midpoint conversion mode is ON, S6
5 will be performed. Here, the input string "defect
The string "skipping" with spaces converted to midpoints "
"Defect Skipping" is a new string to be dictionary matched.

When this string is matched with a dictionary,
According to FIG. 7, "defect skipping" is found, and this word is registered in the dictionary word list (9b).

The process ends through 866, and "defect skipping" is obtained from "defect skipping" as a dictionary lookup result.

FIG. 10 is a diagram for explaining the processing for the character string "delayed pilling". Assume that the space deletion mode and the space←midpoint conversion mode are OFF, and the space-hyphen conversion mode is ON. First, at 860 in FIG. 6, the input character string is matched with the dictionary in FIG. Nothing is set in the word list (log). Then, in S81, it is determined that it is an unknown word, and SB
2. The mode is checked in S64, and the process proceeds to SBB. Then, in 56111, since the space←→hyphen conversion mode is ON, S87 is performed. Here, spaces in the input string "Day-Late Billing" are converted to hyphens, resulting in the string "Day-Late-Pilling".
becomes a new string for dictionary matching. When this character string is matched with the dictionary, "delayed pilling" is found as shown in FIG. 7, and this word is registered in the dictionary word list (10b), and the process ends.

As a result of dictionary lookup, "day-laid pilling" to "
Obtain "Delayed Pilling".

In the above embodiment, spaces in the input character string are either deleted, converted from the midpoint to 1m, or converted to /% in, but spaces can also be converted to symbols other than the above. Possible expansions include replacement, combination of multiple conversion modes, etc.

Further, in the above embodiment, a Japanese-English machine translation system has been described, but by changing the knowledge information stored in the translation dictionary 6, it can be used as an English-Japanese machine translation system. When performing English-Japanese machine translation, the original English text that has been manually translated contains spaces between each word. At this time, the word search process does not stop due to spaces between words. This is because, for example, as shown in FIG.
n, "is input into the English-Japanese machine translation system through the input unit 1. Then, the English sentence is classified by space and each word, that is, "Th1s", "is", "a"
.

The words ``pen'', ``, and ゛ are searched sequentially from the beginning of the sentence and translated to obtain the translated text.Even if you try to input the English sentence shown in Figure 11 and make a typing error, the display screen as shown in Figure 12. Consider the case where an input is made in the original text display area No. 8. Note that the space 1201 is a space due to an input error.

In this case, the English text in Figure 12 is “Th”, “1s”, “IS”.
”, “a”.

Dictionary lookup is performed by classifying each space into "pen", ",".The dictionary used for this dictionary lookup may, for example, store headwords and translations of the dictionary as shown in Figure 13. There is.

Matching of "Th" with the dictionary shown in FIG. 13 is performed. In this matching, "Th" is "th",
In other words, uppercase letters are converted to lowercase letters.

Since there is no headword that matches "th" in the dictionary of FIG. 13, nothing is set in the dictionary word list of FIG. 14 (14g). Therefore, the word is determined to be an unknown word at S61 in FIG. Next, the space 1201 for "This is" is deleted. This is executed by turning on the space deletion mode. As a result, "T old S" becomes a new character string for dictionary matching. When this character string is matched with the dictionary, the first
"Th1s" is found in the dictionary shown in FIG. 3, and this word is registered in the dictionary word list (14b). And “Th1
Next, the word search is performed in the order of the verb "IS" and "a". Note that the last half of "Th1s", "1S", will not be mistakenly recognized as the verb "IS". This is because word searches sequentially search from the beginning of the sentence. Therefore, the space between “th” and “is” is deleted and “this” is searched, and the verb “I” which is the next word after “this” is searched.
S'' is used to perform matching as a single word to see if there is a matching headword in the dictionary.

In addition, there are also titles that contain multiple spaces in the input character string (for example, "Introduction", "
W ORL D”)l;1m, word searches can be performed reliably by the process explained in the above embodiment.Therefore, the operator does not have to take the trouble of correcting the original text or registering all the words in the dictionary. This greatly simplifies the editing work after the translation process is complete, and allows you to easily obtain an appropriate translation result.Also, it is possible to delete the spaces and replace them with symbols. Since the substitution is performed only for unknown words,
Translation processing time will not increase.

[Effects of the Invention] According to the present invention, a word search can be performed reliably even when an unnecessary space is included in a word. In addition, even if there is a discrepancy in the notation, such as a word that is normally displayed with a midpoint or hyphen, but contains a space instead of those symbols, it is now possible to match it with the headword in the translation dictionary. Word searches can be performed reliably.

Therefore, translation processing can be performed efficiently.

[Brief explanation of the drawing]

FIG. 1 is an overall block diagram of an embodiment of the present invention, FIG. 2 is a diagram showing a key arrangement of an input section used in an embodiment of the present invention,
FIG. 3 is a diagram showing an example of the screen layout of the display unit used in an embodiment of the present invention, FIG. FIG. 5 is a flowchart showing an example of the flow of translation processing used in one embodiment of the present invention, FIG. 6 is a flowchart showing the flow of word search processing in dictionary lookup processing used in one embodiment of the present invention, and FIG. , FIG. 13 is a diagram showing the headings and translated words of the translation dictionary used in one embodiment of the present invention,
Figures 8, 9, and 10. FIG. 11, FIG. 12, and FIG. 14 are diagrams for explaining the operation of one embodiment of the present invention. DESCRIPTION OF SYMBOLS 1... Input part, 2... Original text storage part, 3... Translation text storage part, 4... Edit control part, 5... Translation part, 6...
Translation dictionary, 8...display section, 9...printing section

Claims (3)

[Claims] (1) An input unit for inputting the original text to be translated; a translation dictionary unit that stores knowledge information used in translation processing; and a translation process for the original text using the contents of this translation dictionary unit to produce a translated text. and a display unit that displays the original text input from the input unit and the translated text obtained from the translation unit, and the translation unit refers to the translation dictionary unit for the original text to determine the corresponding word. 1. A machine translation system characterized in that when a word search is performed to search for an unknown word in the original text that cannot be searched, a blank space in the unknown word is deleted or replaced with a symbol and the word search is performed again. (2) The machine translation system according to claim 1, wherein the input unit is for instructing whether or not to delete or replace blanks in the unknown word with symbols. (3) The translation unit searches for a set of character strings obtained by removing a blank space between an unknown word that cannot be searched and a character string adjacent thereto in the original text. Machine translation system.