JPS62203265A - Mechanical translation system - Google Patents

Abstract

PURPOSE:To eliminate a trouble in case of command input by the operator to change a dictionary or a grammar used for the translation by giving change command of the kind of a translation dictionary/grammar without intermission of consecutive translation processing before the translation of an optional sentence in plural sentences subject to being translation processing consecutively. CONSTITUTION:In case the necessity of change of a translation dictionary/ grammar to an optional sentence on the way of a consecutive translation processing (conversation translation) exists, the user operates the 'translation dictionary/grammar change key' realized as one of function keys by the user before the translation command of the sentence, then the pattern edit area is displayed. In this example, it is set whether the translation mode (analysis grammar) is the standard analysis grammar or not, whether the translation dictionary (word/phrase dictionary) is a general dictionary or not, whether the style of sentence (generative grammar) is the polite style or not. In using a cursor key or the like, the type of the dictionary/grammar to be changed is instructed and a selection/execution key 120 is operated, then the instructed dictionary/grammar is changed to proceed the succeeding translation processing.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a machine translation system that can appropriately change translation dictionaries and grammar.

[Technical background of the invention and its problems]

In recent years, systems that automatically machine translate input original texts using computers have been attracting attention. This machine translation system basically divides the input original text into predetermined processing units such as morphological and syntactical analysis (words and phrases), and searches the translation dictionary for each processing unit. It is configured to find translated words (translated phrases), etc., and combine them according to predetermined translation rules to obtain the translated text.

However, it is extremely difficult to reliably obtain appropriate translations, partly because the technology for interpreting the meaning of natural languages has not yet been fully established.

Furthermore, the result of translation is not necessarily determined by the exact output for a single sentence, but depends on the type of sentence being translated, the source field, etc., the style of the translated sentence, the type of translated words in the translated sentence, etc.
We need to change fields. Furthermore, when translating long sentences that generally take a long time to translate, it is more time-consuming for the user to correct the rough translation results after outputting them through simple analysis processing than to perform time-consuming and precise analysis processing. In some cases, it may actually be more efficient. These types of translation diversity can be achieved by changing the translation dictionary and grammar used for translation processing.

For this reason, conventionally, before starting continuous translation processing (dialogue translation) for more than - sentences, the user is asked the type of translation dictionary/grammar, has the user specify it, and then the translation dictionary/grammar is created according to the input instructions. It was used uniformly to perform continuous translation processing.

However, the sentences to be serially processed often include sentences of different types and source fields, and short sentences that are easy to translate and long sentences that are not.

In such a case, according to the conventional method, while processing multiple sentences by interactive translation, when the original text for which the translation grammar is to be changed becomes the target of translation, the user must first press the interrupt key to start the translation process. interrupt. Next, the translation completed from the start of translation to the time of interruption is saved in a translation file or the like. Then, the system returns to the many general state before starting the translation process, and instructs the translation dictionary and grammar again. Furthermore, the user inputs the resume key to restart the translation process of the original text after the interruption.

In other words, in the conventional interactive translation format, it is not possible to instruct changes to the translation dictionary or grammar during the translation process (editing the translated text), and if you want to change it, it will return to the initial settings before the translation process started. There was a need. For this reason, the user has the disadvantage that the number of key operations increases, the waiting time increases, and it is not efficient.

[Purpose of the invention]

An object of the present invention is to provide a machine translation system that allows easy editing operations by an operator.

Another object of the present invention is to provide a machine translation machine that can change the dictionary and grammar used for translation during continuous translation work (dialogue translation) by simple instructions from the operator without interrupting the continuous translation work. The goal is to provide a system.

[Summary of the invention]

The present invention is a machine translation system that translates an original text given from an input unit according to a translation processing procedure using knowledge information stored in a translation dictionary, and obtains the translated text by continuously translating a plurality of sentences one by one. When translating a sentence, before starting the translation of a given sentence, the operator inputs the type of translation dictionary and grammar to be used for translation.
The feature is that the translation can be continued by changing the translation.

〔Effect of the invention〕

According to the present invention, it is possible to instruct the type of grammar in the translation dictionary 9 to be changed without interrupting the continuous translation process before translating any sentence among a plurality of sentences that are being continuously translated. Therefore, it is possible to reduce the troublesomeness of the operator when inputting instructions to change the dictionary/grammar used in Translation 1, and to save the time expended from the instruction to the change.

Therefore, editing operations can be significantly simplified and appropriate translation results can be easily obtained.

[Embodiments of the invention]

FIG. 1 is an overall block diagram of an English-Japanese machine translation system as an embodiment of the present invention. This system consists of an input section 11 consisting of a keyboard, an original text storage section 2 that stores the input English text as the original text, a translation storage section 3 that stores the Japanese text corresponding to the input English text as the translated text, and an editing section that controls the entire system. A control unit 4, a translation unit 5 that executes translation processing, and a translation dictionary/grammar 6 that stores knowledge information used in translation processing.
, a display control section 7 that controls the display of the original text, translated text, etc., a display section 8, and a printing section 9 that prints out the original text, translated text, etc. and outputs it as a notebook.

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

The input unit 1 is equipped with the following various keys in addition to character keys for English writing.

Translation instruction keys: 100 Edit keys: 101-106 Function keys: 110-117 Cursor keys: 130-133 Other keys FIG. 3 is a diagram showing an example of the layout of the screen of the display section 8. 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 one-translation processing on the dialogue side 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 S
6, it is monitored whether there is any key input from the input unit 1 or whether a translation completion signal is received from the translation unit 5, and if there is any key input, a corresponding process is performed.

If the operator operates any of the function keys, this is detected in step 85J, and the process corresponding to that function key is performed in step 812.

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

When the operator operates the cursor key 130 or other control keys, processing for moving the cursor corresponding to the key operation and 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). If 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 I [1 section 4 sends the input buffer to the translation section 5. , and instructs the start of the translation process (step 510
). If input editing such as U correction, insertion, deletion, etc. is required during the original text, move the cursor to the desired editing location using the cursor keys, and then press the edit keys such as the insert key 101 or the delete key 102. You can do this using

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

The morphological analysis unit (851) applies rules and
Using the irregular change dictionary 6a, the Z word with a change in the ending etc. is converted into its original form (basic form).

The dictionary search unit (852) searches the word/idiom dictionary 6b for each word constituting the input original text, and obtains information such as its part of speech and translation.

The analysis control unit (853) performs a process of sending translation candidates to the syntactic analysis unit.

The syntax analysis unit (854) uses the analysis grammar 6C to analyze the syntax of the translation candidate and generates an English sentence structure. If the analysis fails, the process returns to the analysis control section.

The structure conversion unit (855) converts the structure of English candy into the structure of Japanese using the conversion grammar 6d. If rejected, return to the parsing section.

The Sentence generation unit (856) determines the Japanese word order from the Japanese structure according to the generative grammar 6e and converts it into a word string.

The morpheme generation unit (857) completes the translated sentence by changing the ending of the word using the morpheme analysis unit 6f.

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

Referring again to FIG. 4 ζ, when the m collection control section 4 detects the translation completion signal in step S1, it proceeds to step 87J and sends the translation candidates and auxiliary information obtained from the translation section 5 to the display control section 7. The translated text is displayed (step 87). The display control unit 7 displays the translation candidate at a position corresponding to the input original text (on the left side of the screen) within the translation display area, and
If there is auxiliary information, the translated word is displayed in high brightness or inverted so that the operator can identify it.

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, by operating the insert key 101, a character is inserted before the cursor position. From the operation number of the delete key 102,
Delete the string pointed to by the cursor. By operating the movement key 103, the range indicated by the cursor is moved. By operating the cancel key 104, the above key 1
Nullifies the effect of 01,102°103. 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, new words and phrases can be registered. By operating the dictionary deletion key 113, it is possible to delete the tsume ψjukusaku registered through dictionary registration. Partial translation key 1
By performing the operation in step 14, a partial translation is displayed when translation fails.

The cursor keys include a cursor movement key 1301 for moving the cursor in each direction, a unit switching key 131 for changing the unit in which the cursor moves, and an inter-area movement key 132 for moving the cursor between display areas.
, and enlargement/reduction keys for reducing the cursor size character by character or word by word.

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 above, the translation dictionary/grammar changing function will be further explained in detail.

As mentioned above, the quality of translation by a machine translation system is determined by the dictionary and grammar used by the system. This dictionary
There are three ways to change the grammar: change the "translation dictionary" (general dictionary or specialized dictionary), and "translation mode".
How to change (standard parsing grammar or simple parsing grammar), "
There are ways to change the "translation style" (is subgrammar or subgrammar), etc. First, I will discuss the necessity of changing translation dictionaries (words and dictionaries).

The word/idiom dictionary 6b describes English words/idioms and their Japanese translations. However, even for the same English word, it may be easier to understand if the translation when used in general sentences is different from the translation when used in sentences in special fields. For example, generally appropriate translations for the word "base" include "foundation,""foundation," and "evidence." However, if the translated text is in the field of mathematics or geometry,
”, “bottom”, etc., and if the translated sentence is in the field of chemistry, “
In many cases, "base" is appropriate. Although it is more specific, in the field of baseball, it is easier to translate it as ``base''. By specifying the field of translation in this way, it is possible to preferentially output translated words specific to that field.

There are big differences not only in the types of translated words, but also in English words, especially those registered as idioms. For example, “f 1o
For the expression ``ppy disk'', the general word dictionary includes ``fIoppy: (adj) fluttering''.

Since only the translated word ``disk: (noun) disk'' is included, it is output as ``flapping disk.'' However, in a specialized dictionary in the computer field, -f 1oppy d
isk: (noun) floppy disk,” it outputs “floppy disk.” In this way, by appropriately switching between the general dictionary and the specialized dictionary depending on the nature of the original text, it is possible to employ words and phrases specific to the field.

Next, we will discuss the necessity of changing the translation mode (parsing grammar).

The analysis grammar 6C describes a procedure for generating an English conceptual structure while examining the parts of speech and word strings of an English input sentence. In a standard syntactic analysis procedure, a conceptual structure is generated by checking an input sentence word by word against an analysis grammar against a series of part-of-speech Φ words from the first word to the last word of the sentence. Since various parts of speech and word sequences are processed using a limited amount of grammar, there are points in the middle of the processing where there is a possibility of branching depending on how the next word is processed. For example, part of the grammar is expressed as shown in FIG.

The analysis of sentences begins from the point of S's al. Processing progresses by following the arc in the direction of the arrow, and processing of the entire sentence ends at point a3 indicated by a double circle.

A process shown in uppercase letters above an arc indicates a step down to another level of grammar with that name, and a process shown in lowercase letters above an arc parses one word that has that name in its part of speech. Show that. An empty arc indicates that no processing is performed and the process proceeds to the next point. or,
When the processing for that level is completed at the point indicated by each double circle,
Generates the structure written on the right side of it, and pops it up at the end of the arc that is pushed down to that grammar. In this example, S represents the entire sentence (Figure 6(a)), and NP represents the noun phrase (Figure 6(a)).
b)), VP(1) is a verb phrase that can have two objects (Figure 6 (C)), and REL is a grammar that analyzes relative pronoun clauses (Figure 6 (C)). d)).

Now, the input sentence ``'I'' give a dog th
When the sentence e man had,' is parsed, the processing proceeds as follows.

■ In the grammar of S, push down to NP to proceed from al to C2.

■ In the grammar of NP, “■” is pronoun (
Since it is a pronoun), proceed from bl to b2ζ, and the seventh body)
Create a structure and pop it up to C2 of S.

■ In the grammar of S, push down to VP to proceed from C2 to C3.

■ VP (in grammar 11, “gave” is ver.
Since it is d (verb), proceed from C1 to C2.

■ To proceed from the same <C2 to 03, NPj is pushed down.

■ In the grammar of NP, "a" is det (determiner).

“dog” is a noun, so from bl to d
The et arc leads to b3, the hairless arc from b3 to b4, and the notn arc from b4 to b5. Since b5 is a double circle, create the structure shown in Figure 7(b) and V P (C3 of 11). I'll upload it to you.

■ Continuing the grammar of VP (1), push down again to NP to proceed from C3 to C4.

■ In the NP grammar, -1he' is det, ”
Since "man" is nourTQ, follow the same arc as ■, create the structure shown in Figure 7 (C) with b5, and create V P (11 (
7) C4 pops up.

■ Since C4 is a double circle, Figure 7 (
b) Using the structure of e (C), create the structure of Figure 7(d) and pop it up at 83 of S.

@C3 finally generates the order of FIG. 7(e) from FIG. 7(a) and FIG. 7(d), and ends the process.

However, with this method, the last word "had" of the input sentence is dropped, resulting in a failure in parsing the sentence. So, considering the processing up to that point, we go back to the point where there is a possibility of another processing, and start again. 7(f) is generated. Then, pop up to b6 of NP, make Figure 7 (g) from Figure 7 (b) thorn), pop up to VP (03 of 11, and then pop up from C3 to C3 of S, and finally FIG. 7 (The structure of hl is generated and the analysis of the - sentence is successfully completed.

In this way, it often happens that the analysis fails midway through the analysis, returns to a point where another process is possible, and reanalyzes.
This increases time and memory consumption.

On the other hand, suppose that only simple parsing grammars are allowed to be output as a collection of partial parsing results. For this purpose, consider reducing the grammar to Figures 8) to (C). In this way, the processing of S is always completed up to "I give a dog", creating the structure shown in FIG. 9(a).

As a continuation, "the man had," is processed as another S, creating the structure shown in FIG. 9(b). In this case, each analysis is shorter than a regular analysis, so there is less chance of failure and retrying, and even if you do redo it, the scale will be small, so less time and memory will be consumed.

Furthermore, I will discuss the necessity of changing the translation style (generative grammar, morpheme generative grammar).

The generative grammar 6e and the morpheme generative grammar 6f describe a procedure for generating a Japanese translation from the conceptual structure of the Japanese language.

In the case of Japanese, there are two styles of writing: ``desu・mazu'' and ``deru'', but for letter translations, ``desu・mazu'' is appropriate, and for technical sentences, ``desu'' is appropriate. There are many cases.

Such a difference in writing style can be realized by changing the generative grammar 6e and the morpheme generative grammar 6f.

Based on the above three types of change methods, continuous translation processing (
We will explain the process of changing the translation dictionary and grammar during dialogue translation.

If it is necessary to change the translation dictionary or grammar for a given sentence during continuous translation processing (dialogue translation), the user can use one of the function keys to change the translation dictionary or grammar for a given sentence. When the "translation dictionary/grammar change key" is operated, the display shown in FIG. 10 is displayed in the editing area of the screen. In this example, the translation mode (analysis grammar) is set to standard analysis grammar, the translation dictionary (word φ compound dictionary) is set to a general dictionary, and the translation style (generative grammar) is set to . Here, by using the cursor keys or the like to specify the type of dictionary/grammar to be changed and operating the selection/execution key 120, the specified word '4e grammar is changed and subsequent translation processing is continued.

FIG. 11 is a flowchart of the editing control section for changing the translation dictionary/grammar according to the dialogue translation. FIG. 11(a) is a diagram showing the process from the start to the end of dialogue translation. When the user specifies a translation dictionary/grammar on the menu shown in FIG. 10 (step 14), the initial state of the translation dictionary/grammar is set according to the instruction (step 515).

Then the actual dialogue translation begins (step 16).
The details are as shown in FIG.

FIG. 11(b) is a flowchart for changing the translation dictionary and grammar during dialogue translation. This figure is a detailed view of steps 85 and subsequent steps in FIG. 4. When a function key input is accepted (step 5), it is determined whether the input is a "dictionary key" (translation dictionary/grammar change key) (step 17). If the "dictionary key" is input, the menu shown in FIG. 10 is displayed, and the user changes the translation dictionary/grammar designation (step 18).

In this way, it can be specified before the interactive translation process (
Similar to step 15), the specification can be changed even during dialog translation (step 18). The processes of steps 15 and 18 are internally equivalent and will be described below.

FIGS. 11(C) to 11(f) are flowcharts showing the flow of processing after "translation dictionary/grammar change" is instructed. In the example shown in FIG. 10, there are three types of changes, so they are processed as follows.

The editing control unit determines whether there is an instruction to change the "translation mode" (step 20), and if so, performs the "analysis grammar change" process. Therefore, judge the content of the instruction (Step 21). If the instruction is "Standard", set the "Standard Parsing Grammar" to the parsing grammar (Step 22); if the instruction is "Simple", make the "Simple Parsing Grammar" the parsing grammar (Step 23). ). (Figure 11 (CHd
)) Next, it is determined whether there is an instruction to change the "translation dictionary" (step 30), and if so, the "word/jukuame jiyo change" process is performed. Therefore, judge the content of the instruction (step 31). If the instruction is "general", change only the "general dictionary" to the translation dictionary 1iF (word/idiom dictionary) step 32); if the instruction is "general specialized", select the "general dictionary". and convert the "specialized dictionary" into a translation dictionary (step 33). (FIG. 11(C)(e)) Furthermore, it is determined whether there is an instruction to change the ``translation style'' (step 40), and if so, the ``generating grammar change'' process is performed. Therefore, judge the content of the instruction (step 41), and if it is one instruction, make the sub-grammar ``is'' a generative grammar (step 42).
, if it is a "desu/masu" instruction, the "desu/masu subgrammar" is made into a generative grammar (step 43). (Figure 11(e)(f)
)) According to the present invention, in the middle of interactive translation in which multiple sentences are successively translated, the translation dictionary and You can change the type of grammar.

Therefore, even if a continuous sentence consists of multiple sentences, sentences from various types, sources, and fields are mixed together, long sentences that are difficult to analyze, and short sentences that are easy to analyze are mixed,
In continuous translation processing (dialogue translation), it is possible to specify the appropriate translation dictionary and grammar for each sentence, eliminating interruptions in translation work and wasting time due to changes in dictionary and grammar. I can do it.

Therefore, it is possible to provide a machine translation system in which continuous translation processing by an operator is easy, and a great practical effect can be achieved.

Note that the present invention is not limited to the embodiments described above. Although translation from English to Japanese has been described here, translation may be performed in the opposite direction, and the present invention can be similarly applied to translation processing between other languages.

Further, the translation dictionary/grammar may be changed by a method other than displaying it in the editing area and selecting it with a cursor. For example, "translation mode", "translation dictionary".

A unique function key may be provided for each "translation style", and each dictionary/grammar may be changed each time the key is pressed.

Furthermore, the range of translation dictionaries and grammars to be changed is as follows:
It is not limited to "word/idiom dictionary", "analytical grammar", and "generative grammar". "Transforming English conceptual structure into Japanese conceptual structure"
Among the conversion grammars, complex structural conversions take a relatively long time to process, while those that require simple attribute changes can be processed in a short time. Therefore, a standard conversion mode that realizes all conversion grammars and a simple standard mode that only realizes simple conversion grammars that do not take much time may be provided, and the operator may be instructed to select one. Conversely, it is also possible to limit what the operator can select to only "generative grammar."

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

[Brief explanation of drawings]

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. Figure 5 is a flowchart showing an example of the flow of translation processing used in an embodiment of the present invention, Figure 6 is a diagram showing an example of syntactic analysis grammar in regular translation, and Figure 7 is a flowchart showing an example of the syntactic analysis grammar used in an embodiment of the present invention. Figure 8 is a diagram showing an example of syntactic analysis grammar in simple translation. Figure 9 is an internal data structure when analyzed using the grammar in Figure 8. Figure 10 is a diagram showing an example of the display when instructing to change the translation dictionary Kaoru Grammar.
The figure is a diagram showing an example of the flow of processing when a translation dictionary-grammar change is instructed. 1: Input section 2: Original text storage section 3: Translation storage section 4: Editing control section 5: Translation section 6: Translation dictionary 8: Display section 9: Intobe agent Patent attorney Nori Chika Ken Yudo Takehana Kikuoji Shi 〇- (1)? Bow D -no
1) (1) (1) (1) η Q(a) (
b) (C) <t> (Cl) No. 7 (a) No. C] Figure 8 (b) and Figure 9

Claims (1)

[Claims] An input section that has at least character keys, edit keys, and cursor keys, an original text storage section that stores the original text to be translated inputted from this input section, and knowledge information such as a word and phrase dictionary and grammar used for translation processing. The translation and dictionary department accommodated
a translation unit that executes translation processing of the input original text using the contents of the translation dictionary unit; a translation storage unit that stores the translated text obtained as a result of the translation processing; and a display for displaying the original text and the translated text. and a translation for controlling these and for editing the original or translated text displayed on the display section and controlling the cursor on the display surface of the display section in response to the operation of the edit key or cursor key. an editing section, the translation editing section changing the contents of the translation dictionary section in response to an operation of an edit key that instructs to change the translation dictionary during editing processing.