JPH10254881A - Machine translation device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To search a part corresponding to variable meaning origin from a sentence constituent element by finding a matching degree by setting a character string to be that matched with the variable meaning origin and converting the character string into a translation word registered in a word dictionary. SOLUTION: The analysis part 10 of a CPU 1 calculates the matching degree of a selected fixed sentence with an input sentence. The translated sentence matched with the fixed sentence is stored in a work memory 5. A variable meaning attribute element in the Japanese sentence of the fixed sentence stored in the work memory 5 is substituted for the variable meaning origin element in the input sentence. The variable meaning origin element in an English sentence is substituted for the translation word of the element stored in the work memory 5. When the sentence constituent elements except the variable meaning origin are almost matched and the variable meaning origins are not matched, the translated sentence of the character string equivalent to the variable meaning origin is extracted from the word dictionary 19 and a sentence approximated to the sentence which a user inputs and the translated sentence is securely displayed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine translator for translating a conversation sentence in a first language into a second language using a pre-registered conversational example sentence.

[0002]

2. Description of the Related Art At the present time when the number of opportunities to go abroad for sightseeing and the like has increased, many book examples of conversations that collect many conversational expressions have been published so that conversations in foreign countries are not troubled. Also,
There has been proposed a machine translation apparatus that digitizes these conversation examples and selects a conversation example by key operation or the like and displays a translation of the selected conversation example. For example, Japanese Unexamined Patent Publication
The invention of -325072 corresponds to this.

In this machine translation apparatus, a variable section is provided in advance in a conversational example sentence, and the same example sentence is used even when a part of words of an input free sentence is replaced.
It is possible to increase the variations of example sentences while saving memory.

Further, there has been proposed a machine translator which automatically selects an optimal example sentence for conversation by freely inputting a sentence to be translated and displays the translated sentence. This machine translator also has a variable unit, and by automatically replacing the input phrase, a sentence close to the input sentence is generated.
It is configured to output.

In this method, a large number of conversational example sentences and phrases characterizing each example sentence grouped by synonyms are stored. The former is called a fixed phrase, and the latter is called a semantic feature. A variable semantic feature is called a variable semantic feature.

In this method, semantic features are extracted from an input sentence, and a fixed sentence having a pattern of a combination thereof is searched. At this time, it is necessary to determine how much the input sentence and each fixed sentence match as a pattern of a combination of semantic features.

The prior art employs a method in which the degree of coincidence between an input sentence and each fixed sentence is calculated by matching each word unit, and only fixed sentence having a certain threshold or more is displayed.

[0008]

However, if the input sentence and a certain fixed sentence do not match only in the variable semantic feature and all other parts of the sentence match, the conventional method calculates the degree of match. However, there is a case where the machine does not display the image even though the degree of coincidence does not exceed the threshold value and the person intuitively senses the meaning close.

[0009] As an example, suppose that there is an input of "I was separated from the tour operator." On the other hand, it is assumed that there is no fixed sentence having the same sentence, and the fixed sentence “I was separated from a friend” is the closest sentence. Here, it is assumed that the part corresponding to “friend” is a variable semantic feature, and the feature “passenger” is not registered.

According to the present invention, when sentence constituent elements other than the variable semantic features almost coincide with each other and the variable semantic features do not coincide with each other, words included in a word dictionary incorporated in the apparatus and their translations are used. By synthesizing a bilingual sentence, a fixed sentence whose variable semantic features do not match and its translated sentence are displayed. When the word dictionary is used for replacing the variable semantic feature of a fixed sentence in this way, it is necessary to find out which part of the sentence component corresponds to the variable semantic feature.

As an example, "I missed the train station at the station."
Suppose there is an input sentence On the other hand, it is assumed that there is no fixed sentence having the same sentence, and the fixed sentence “I was separated from a friend” is the closest sentence. here,
It is assumed that the part corresponding to “friend” is a variable semantic feature, and the feature “passenger” is not registered. At this time, it is necessary to search for a character string that can be replaced with “friend” from the character string “at the tour operator and the station” that does not match the fixed phrase.

In the present invention, when a word dictionary is used for replacing a variable semantic feature of a fixed sentence, a set of character strings that can be connected immediately after the variable semantic feature is defined, so that variable sentence components can be used to define variable semantics. Find the part corresponding to the feature.

Further, even if a sentence element corresponding to the variable semantic attribute is narrowed down by using a set of character strings that can be connected immediately after the variable semantic feature, the word dictionary is added to the character string that does not match the fixed phrase. There may be multiple registered words.

As an example, suppose there is an input sentence "Do you have plums?" On the other hand, it is assumed that there is no fixed sentence with the same sentence, and the fixed sentence “Do you have an apple” is the closest sentence. Here, it is assumed that the portion corresponding to "apple" is a variable semantic feature, and neither the feature "peach" nor the feature "sumo" is registered. Further, it is assumed that both the word “sumo” and the word “peach” are registered in the word dictionary. At this time, it is necessary to search for a partial character string that can be replaced with "apple" from the character string "sumomo" that does not match the fixed phrase.

In the present invention, a word dictionary is used for replacing variable semantic features of a fixed phrase, and when a plurality of words registered in the word dictionary exist in a character string that does not match the fixed phrase,
By selecting the longest character string, a part corresponding to the variable semantic feature is searched from the sentence components.

Also, by using a set of character strings that can be connected immediately after the variable semantic feature, the sentence components corresponding to the variable semantic attributes can be narrowed down or the length of the character string registered in the word dictionary can be reduced. Even when narrowing down to only long ones, there are cases where a plurality of candidate portions corresponding to variable semantic features exist in a character string that does not match the fixed phrase.

As an example, assume that there is an input sentence "Please carefully pack in a wooden box". On the other hand, it is assumed that there is no fixed sentence with the same sentence, and the fixed sentence “Please pack in a case” is the closest sentence.
Here, it is assumed that the portion corresponding to “case” is a variable semantic feature, and the feature “wooden box” is not registered.

Further, it is assumed that the word "wood box" and the word "polite" are registered in the word dictionary, and the word class of "wood box" is a noun and the word class of "policy" is an adjective verb. . At this time, it is necessary to search for a partial character string that can be replaced with “case” from a character string “Carefully in a wooden box” that does not match the fixed phrase.

In the present invention, when the word dictionary is used for replacing the variable semantic feature of a fixed sentence, the part of speech of the variable semantic feature is registered and registered in the word dictionary in a character string that does not match the fixed sentence. When there are a plurality of words, a part corresponding to the variable semantic feature is searched for from the sentence constituent elements by selecting a word having the same part of speech from the words in the word dictionary.

[0020]

[Means for Solving the Problems]

<Means of Claim 1> A bilingual database in which example sentences of a conversation in a first language, sentences in a second language which are translations thereof, and semantic features are registered as a pair, a semantic feature and its semantic features Input a sentence for retrieving the bilingual database, a semantic feature dictionary that defines and stores a notation to be extracted as a variable, and a variable semantic feature dictionary that has the same semantic feature and replaceable words are registered. Input means, search means for searching the bilingual database, the semantic feature dictionary and the variable semantic feature dictionary based on the input sentence input from the input means, and the bilingual database and the semantic feature by the search means Replacement means for replacing the example sentence searched based on the dictionary based on the semantic variable semantic feature dictionary; and an example sentence extracted by the search means and the replacement means Analysis means for calculating the degree of coincidence with the input sentence input from the input means; output means for outputting a translation of an example sentence based on the analysis means;
A word dictionary capable of searching for a translation of a second language from a word of a language, wherein the analyzing means includes, in the input sentence, the semantic feature dictionary and the variable semantic feature dictionary as the semantic features. If there is a character string that is not registered in the word dictionary and is registered in the word dictionary, the degree of matching is determined assuming that the character string matches the variable semantic feature. The above problem is solved by converting the word into a translation registered in the word dictionary.

<Means of Claim 2> In the bilingual database, a connectable character string is added after the character string having each semantic feature, and the analyzing means includes the semantic feature as the semantic feature in the input sentence. If there is a character string that is not registered in the feature dictionary and the variable semantic feature dictionary but is registered in the word dictionary, a character string that can be connected immediately after the semantic feature dictionary and a character string If the immediately following character string matches, the degree of matching is determined assuming that the character string matches the variable semantic feature, and the replacing means converts the character string into a translated word registered in the word dictionary. Solve the problem.

<Means of Claim 3> In the analysis means, the input sentence is not registered as the semantic feature in the semantic feature dictionary and the variable semantic feature dictionary.
For the longest character string among the character strings registered in the word dictionary, the degree of coincidence is determined as that matching the variable semantic feature, and the character string is registered in the word dictionary by the replacing means. The above problem is solved by converting to a translated word.

<Means of Claim 4> In the word dictionary, in addition to the words in the first language and the translations in the second language, the parts of speech of the words in the first language are registered, and the respective meanings are registered in the bilingual database. The part of speech of the feature is registered, and in the analysis unit, the input sentence is not registered as the semantic feature in the semantic feature dictionary and the variable meaning feature dictionary, but is registered in the word dictionary. A character string whose part of speech matches the part of speech of the semantic feature is determined as a character string that matches the variable semantic feature, and the degree of matching is determined, and the replacement unit converts the character string into a translated word registered in the word dictionary. Solves the above problem.

[0024]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to the drawings. In the following example, Japanese is used as the first language and English is used as the second language. It is not particularly limited to these languages.

FIG. 2 shows the appearance of a machine translation apparatus according to an embodiment of the present invention. Many pairs of Japanese and English sentence examples are stored, and these can be displayed on the liquid crystal display 2. In addition, this device uses handwritten character recognition using a dedicated pen 15 as an interface. A display other than liquid crystal such as a CRT can be used, and other means such as voice or keyboard input can be used as input means.

FIG. 1 shows an outline of a circuit configuration of such a machine translation apparatus. The machine translator is a CPU (central processing unit) 1 for performing a central function of various controls.
Has been arranged. The CPU 1 sends the ROM 6, the working memory 5, the display driver 3, and the like through a bus 18 such as a data bus.
It is connected to each of the input recognition units 4.

The ROM 6 is a large-capacity read-only memory that stores a program for performing various controls of the machine translation device, a large number of phrases, and the like. The working memory 5 is a random access memory for temporarily storing data required for executing the program. The display driver 3 is a driver for displaying a bilingual sentence or the like on the liquid crystal display 2. The input recognition unit 4 recognizes input data from the handwriting input board 14 in FIG.

In the ROM 6, a bilingual database 7, a semantic feature dictionary 8, a variable semantic feature dictionary 9, a word dictionary 1
9 is stored. The bilingual database 7 has addresses 0 to 50,000, for example, and the semantic feature dictionary 8 has addresses 50001 to 7
0000, the variable meaning feature dictionary 9 has addresses 70001 to 10
0000, word dictionary 19 is 100000-3000 addresses
00 is stored.

The CPU 1 is divided into an analysis unit 10, a search unit 11, a replacement unit 12, and a control unit 13 according to the nature of the operation. The analysis unit 10 analyzes the input data recognized by the input recognition unit. The search unit 11 searches the bilingual database 7, the semantic feature dictionary 8, and the variable semantic feature dictionary 9 stored in the ROM 6. The replacement unit 12 replaces a variable semantic feature element.
The control unit 13 performs control other than the above, such as control of screen display.

An embodiment of the data structure of the bilingual database 7, the semantic feature dictionary 8, the variable semantic feature dictionary 9, and the word dictionary 19 will be described below. These are common in claims 1 to 4 except for a part.

The bilingual database 7 defines a number of useful example sentences in the following form that are desired to be translated during overseas travel and transmitted to local people.

Example 1 How much is it to New York? How much is it to New York? How much is the place name? Example 2 How much is New York to Chicago? How much is it from New York ｝ $ 1 $ to ｛Chicago｝
$ 2 $? How much is from ｛地名｝ to ｛地名｝?

Here, the first line of the sentence example is the Japanese sentence to be displayed, the second line is the translated sentence, the English sentence displayed together with the Japanese sentence as a translation result, and the third line is the corresponding translated sentence. It is a set of semantic features that This is called a semantic feature sequence. Also, the portion surrounded by ｛｝ is a variable semantic feature.

The semantic feature is a character string serving as a keyword when searching for a sentence example, and is often assigned one to one sentence of each Japanese sentence. In “How much is it up to New York” in Example 1, “New York” and “until”
"How much" and "how much" are the semantic features of this sentence.
To be precise, these are the names of semantic features, and what character strings are defined as these semantic features are described in the semantic feature dictionary 8.

Further, "New York" in the sentence example 1 is a variable semantic feature. If various places such as "London" and "Oxford" are included in this part of the sentence example 1, the number of sentences that can be translated increases, which is very effective. Therefore, a Japanese character string that can enter this location and its translation are defined in the variable meaning feature dictionary 9, and the set is treated as one variable meaning feature called "place name".

In the bilingual database 7, the name of the variable semantic feature is described by enclosing it in the semantic feature sequence part, and one element thereof is written in parentheses in the Japanese sentence on the first line. Also, write the translated word in parentheses in the English sentence on the second line. Japanese text and English text are defined as parallel translations in the variable meaning feature dictionary 9. These also serve as representative values of the variable semantic feature in this sentence. That is, if there is no character string corresponding to the variable semantic feature in the input, this element is displayed.

For example, sentence example 1 has one variable semantic feature named "place name", the representative value of the variable semantic feature in this sentence is "New York", and its translation is "NewYork". If there are a plurality of variable semantic features in one sentence, the order of appearance of the variable semantic features in the Japanese sentence matches the order of appearance of the variable semantic features in the semantic feature sequence.

Similarly, sentence example 2 has two variable semantic features named “place name”, and the representative value of the first “place name” variable semantic feature in this sentence is “New York”.
The translation is "New York." The representative value of the second "place name" variable semantic feature is "Chicago", and its translation is "Chicago".

Also, $ 1 $, $ 2 $ after the $ symbol in sentence 2
Indicates that the “place name” variable semantic attribute corresponds to the first and second variable semantic features in Japanese sentences, respectively.

That is, the first “place name” variable semantic feature containing “New York” is {New York} $ 1 $,
The second "place name" variable semantic feature containing "Chicago" corresponds to "Chicago" $ 2 $. When a semantic feature that is not variable is distinguished from a variable semantic feature, it is called an invariant semantic feature.

In the semantic feature dictionary 8, names of semantic features and Japanese character strings to be extracted as the semantic features are defined. For example, "from""to""howmuch""isit?"
The semantic feature is defined in the semantic feature dictionary 8 as follows, from "[Semantic feature dictionary 8]" to "".

This means that the Japanese character string on the right is defined as the semantic feature of the name on the left.
For example, the character string "how much" is defined as a semantic feature of the name "how much".

In the variable semantic feature dictionary 9, the name of the variable semantic feature, a Japanese character string as an element included therein,
The corresponding English translation is defined. For example, a set of elements of the variable semantic feature “place name” and its translations is defined in the variable semantic feature dictionary 9 as follows: [variable semantic feature dictionary 9] place name New York Chicago Chicago London London Oxford Oxford .

In this method, the Japanese character string on the left side of the second and subsequent lines (eg, “New York”) is defined as an element of a variable semantic feature “place name”, and its translated word is the English character string on the right side of the character string. (Such as "New York").

In the word dictionary 19, Japanese headwords, parts of speech, and English translations are defined. For example, the parts of speech and translations of the words "school" and "tourist" are defined in the word dictionary 19 as [word dictionary 19] school noun school tourman noun a tour conductor.

This means that, for a Japanese character string of school, the part of speech is a noun and the English translation is school. It is not necessary to newly create a word dictionary for the machine translation apparatus, and an existing word dictionary or electronic dictionary may be used.

Hereinafter, the operation of the machine translation apparatus will be described. The following is a description of the processing common to the claims. Embodiments of the individual claims are described below. The outline of the basic operation of this machine translator is as follows. A user inputs a translated sentence, detects a semantic feature from the sentence, detects a fixed sentence from the semantic feature, replaces and displays a variable semantic feature in the fixed sentence. To do. Hereinafter, the translated sentence input by the user is the input sentence,
Japanese sentences stored in the machine translator are called fixed sentences. When a pair of a fixed sentence and its translated English sentence is referred to, it is called a bilingual sentence.

The operation of the machine translation apparatus will be described with reference to the flowchart of FIG. The description will be given in a form showing details of each step of the flowchart of FIG. Take the process of translating the sentence "How much is it up to London" as an example (hereinafter this sentence is referred to as input sentence 1).

<Step A1> Controller 13 of CPU 1
Displays the “translated sentence input screen” of FIG. 4 on the liquid crystal display 14 through the display driver 3.

<Step A2> In the translated sentence input screen shown in FIG. 4, the user uses the special pen 15 to manually input “How much is it up to London” on the handwriting input board 16? FIG. 5 is a diagram in the middle of the input, and FIG. 6 is a state in which the input is completed.

<Step A3> The user presses the interpreter icon 17 in FIG. This input sentence is recognized by the input recognition unit 4 and stored in the work memory 5. Thereafter, the process proceeds to step A4.

<Step A4> Analysis section 10 of CPU 1
Is the input sentence 1 stored in the working memory 5 in step A3.
Then, a partial character string is created and stored in the working memory 5. Next, the process proceeds to step A5.

The partial character strings of the input sentence 1 are as follows. Loron Rondo London London or London: How much: How much: These are stored in the working memory 5.

<Step A5> Search unit 11 of CPU 1
Searches the semantic feature dictionary 8 and the variable semantic feature dictionary 9 stored in the ROM 6 using the partial character string of the input sentence 1 stored in the working memory 5 in step A4 as a key.
If there is a character string defined as a semantic feature in the semantic feature dictionary 8 among the partial character strings, the search unit 11 of the CPU 1 stores the character string and the name of the semantic feature in the working memory 5 as a pair, and proceeds to step A6. Move on to If not, the process proceeds to step A16. At this time, the substrings are searched in order from the front,
By storing the order, the order of appearance in the input sentence of each semantic feature is also stored.

The process of searching the semantic feature dictionary 8 and the variable semantic feature dictionary 9 using the partial character string "how much" of the input sentence 1 as a key will be described below. The search unit 11 of the CPU 1 has a ROM
6 is read from the semantic feature dictionary 8 in order from address 50001, one line at a time, and stored in the working memory 5.
Each row of the semantic feature dictionary 8 has the name of each semantic feature on the left side,
The character string defined as the semantic feature is described on the right side.

It is assumed that [semantic feature dictionary 8] is stored in semantic feature dictionary 8 in this manner.

Now, the character string “how much” is stored in the semantic feature dictionary 8
Therefore, the search unit 11 of the CPU 1 stores the semantic feature name “how much” on the left side and the partial character string “how much” of the input sentence in the working memory 5 as a pair.
Since the character string "how much" is not defined in the semantic feature dictionary 8 in other places, the search of the semantic feature dictionary 8 is terminated.

Next, the search unit 11 of the CPU 1 searches the variable semantic feature dictionary 9 using "how much" as a key. Since the character string "how much" does not appear in the variable semantic feature dictionary 9, the search using "how much" as a key ends, and the next partial character string is set as a key.

The process of searching the semantic feature dictionary 8 and the variable semantic feature dictionary 9 using "London" as a key will be described below. Now, assume that the character string “London” is stored in the working memory 5.

The search unit 11 of the CPU 1 checks whether a character string “London” is defined in the semantic feature dictionary 8 stored in the ROM 6. Now, since the character string “London” does not appear in the semantic feature dictionary 8,
Next, the variable semantic feature dictionary 9 is searched.

The variable semantic feature dictionary 9 has the following description: [variable semantic feature dictionary 9] Place name New York Chicago Chicago London London Oxford Oxford

The character string “London” is a variable semantic feature dictionary 9
Therefore, the search unit 11 of the CPU 1 stores the name “place name” of the variable semantic feature including the name and the translated word “London” of the element in the work memory 5. Since the character string "London" is not defined elsewhere in the variable meaning feature dictionary 9, the search of the variable meaning feature dictionary 9 is terminated. The above operation is performed for all partial character strings of the input sentence 1. It is assumed that the following variable / invariant semantic features are detected from the input sentence 1.

[Semantic features corresponding to partial character strings] "London" = "place name": variable semantic feature "to" = "to": invariant semantic feature "How much" = "how much": invariant semantic feature "is" = “Is”: Invariant semantic feature Here, the right side of = is the name of the semantic feature, and the character string on the left side is detected as a variable / invariant semantic feature of the name on the right side.

Further, the translation “London” of the variable semantic feature element “London” is also stored. [Variable semantic feature element] "London" = "London" Here, the right side of = is a translation of the relevant element of the variable semantic feature, and means that the right side phrase is detected as a translation of the left side phrase, respectively.

Each semantic feature detected from the input sentence is assumed to be detected in the order from the top to the bottom of the correspondence table. This order is stored in the working memory 5. In this example, since the semantic feature was searched, the next step A6
Move on to

<Step A6> Search unit 11 of CPU 1
Searches the bilingual database 7 stored in the ROM 6 using the name of the variable / invariant semantic feature stored in the working memory 5 in step A5 as a key. If any of the variable / invariant semantic feature names appear in the bilingual database 7, the fixed form sentence in which the semantic feature appears is stored in the working memory 5 together with the English sentence of the translated sentence and the semantic feature sequence. Move to A7. If not, the process proceeds to step A16. Taking the input sentence 1 as an example, the details of step A6 will be described.

First, the CPU 1 searches the bilingual database 7 using one of the semantic feature sets stored in the work memory 5 as a key in step A5. First, the bilingual database 7 is searched using the semantic feature “place name” as a key. Now, as shown below, the bilingual database 7 Suppose there is such a description.

In sentence example 1, the “sequence name” semantic feature is included in the semantic feature sequence. Therefore, the search unit 11 of the CPU 1 stores the sentence example 1 in the working memory 5. In the bilingual database 7, other places do not include the “place name” semantic feature. Thus, the search using “place name” as a key is terminated.

The search unit 11 of the CPU 1 performs the above-mentioned operation for all semantic features stored in the work memory 5. Now, it is assumed that the following sentence examples 1 and 3 are retrieved and stored in the working memory 5 after the search for all semantic features is completed.

[0070] Sentence example 3 is searched by the semantic feature “how much”. The above is the description of Step A6 according to the example. In this example, since a fixed phrase has been searched, the process proceeds to step A7.

<Step A7> Analysis section 10 of CPU 1
Selects one sentence stored at the head among the fixed sentences stored in the working memory 5 in step A6. The selected fixed phrase is flagged. Here, sentence example 1 is selected and a flag is set.

<Step A8> Analysis section 10 of CPU 1
Calculates the degree of coincidence between the fixed sentence selected in step A7 or A12 and the input sentence. Here, the calculation is performed using the number of the semantic features of the fixed sentence that match the semantic features in the input sentence. Then, the pair of the fixed phrase and the degree of coincidence is stored in the working memory 5, and if the degree of coincidence is equal to or more than a certain threshold, the process proceeds to step A9.
Move to 3. Here, the threshold is set to 200 points.

Now, sentence example 1 is selected. Semantic features of sentence example 1 "place name""until""howmuch""isit?"
Are all in the input sentence. The number of matching sentences is multiplied by 100, and the matching degree between the fixed sentence and the input sentence is set to 400 points. Therefore, the sentence example 1 and the matching score of 400 are stored in the working memory 5. Since the matching score of 400 in the example 1 exceeds the threshold value of 200, the process proceeds to step A9.

<Step A9> Analysis section 10 of CPU 1
Determines whether the fixed phrase stored in the working memory 5 in step A8 has a variable semantic feature. If there is, the process proceeds to step A10, and if not, the process proceeds to step A11.
Since sentence example 1 has a variable semantic feature "place name", step A
Move to 10.

<Step A10> Replacement unit 1 of CPU1
2 replaces the variable meaning attribute element in the Japanese sentence of the fixed sentence stored in the working memory 5 with the variable meaning feature element in the input sentence 1. The variable semantic feature element in the English sentence is replaced with a translation of the element stored in the working memory 5. The bilingual sentence is stored in the working memory 5 in the replaced form. afterwards,
Move to step A11. Now, the element “New York” is included in the “place name” variable semantic feature of the sentence example 1. on the other hand,
The “place name” variable semantic feature element detected from the input sentence 1 is “London”. At the same time, the translation of "London"
“ndon” is also stored in the working memory 5. Therefore, C
The replacement unit 12 of PU1 replaces "New York" with "London" in Example 1, and replaces "New York" with "London". It looks like this:

[0076] The data is stored in the working memory 5 in this manner, and the process proceeds to Step A11.

<Step A11> Analysis section 1 of CPU1
A value of 0 checks whether there is a fixed sentence with no flag in the work memory 5. Step A12 if present
If not, the process proceeds to step A14. Since no flag is set in the sentence example 3, the process proceeds to step A12.

<Step A12> Analysis section 1 of CPU 1
"0" selects one sentence in the working memory 5 with no flag set and sets a flag. Then, the process proceeds to Step A8. Here, sentence example 3 is selected, a flag is set, and the routine proceeds to step A8.

<Step A8> Analysis section 10 of CPU 1
Calculates the degree of coincidence of Sentence Example 3. Among the semantic features of Sentence Example 3, "How much" is in the input sentence
"Is it?" The remaining “today”, “exchange”, and “rate” do not exist in the input sentence. Further, the two semantic features “place name” and “until” in the input sentence do not exist in sentence example 3. Therefore, minus 50, which is obtained by multiplying the number of matching 2 by 100 and subtracting the number of unmatching 5 by 50, is defined as the matching degree of the fixed sentence with the input sentence.
Therefore, the sentence example 2 and the degree of coincidence of −50 are stored in the working memory 5. Since the matching score of −50 in Sentence Example 3 is lower than the threshold value of 200, the process proceeds to Step A13.

<Step A13> Analysis section 1 of CPU1
A value of 0 deletes from the working memory 5 the fixed phrase stored in the working memory 5 in step A8 and whose degree of coincidence is less than the threshold value. Then, the process proceeds to step A11. Now, the sentence example 3 having a matching degree of -50 points is deleted from the working memory 5,
Move to step A11.

<Step A11> Analysis section 1 of CPU1
A value of 0 indicates that there is no fixed phrase with no flag set in the working memory 5, and the process proceeds to step A14.

<Step A14> Control unit 1 of CPU1
3 is a liquid crystal display 14 through the display driver 3
Is switched from the free text interpretation input screen of FIG. 6 to the bilingual text display screen of FIG. Thereafter, the process proceeds to step A15.

<Step A15> Control unit 1 of CPU1
Reference numeral 3 denotes a liquid crystal display 14 for displaying Japanese and English sentences stored in the working memory 5 through the display driver 3.
Is displayed in parallel as shown in FIG. At this time, the variable semantic features are underlined and displayed. In addition, when displaying, fixed phrases with a high degree of matching are displayed in order. Now, only the sentence example 1 is stored in the working memory 5. The Japanese and English sentences are displayed as shown in FIG.

The description of the operation of the machine translator using the input sentence 1 as an example is now completed. Next, FIG. 3 which was not traced above
Steps in the flowchart will be described. If no semantic feature or fixed form sentence is found in step A5 or A6, it is displayed as translatable. That process is step A16 below.

<Step A16> Control unit 1 of CPU1
3 is a liquid crystal display 14 through the display driver 3
Is switched from the free text interpretation input screen of FIG. 6 to the non-translatable screen of FIG. The above is the common operation of the machine translation device of the present invention.

In this embodiment, the retrieved sentence is displayed as a translation, but only a Japanese sentence or an English sentence may be displayed. Also,
4 to 9 are examples of screen designs to be displayed,
Various designs are conceivable depending on the application. Further, in the present embodiment, as the degree of coincidence between the input sentence and the fixed sentence, a number obtained by multiplying the number of matching semantic features by 100 and subtracting the number of non-matching semantic features by 50 is subtracted. In addition to the above, various methods such as dividing by the number of semantic features in the input sentence to obtain a relative evaluation of 100 points or more can be considered.

Hereinafter, embodiments will be described for each claim. In each embodiment, as in the above-described basic embodiment, a process of displaying a bilingual sentence for a certain input sentence will be described as an example. Hereinafter, an embodiment of the machine translation device according to claim 1 will be described. The process of translating an input sentence by the machine translation device according to claim 1 is the same as in the basic embodiment, except for step A8. Therefore, only the difference in step A8 will be described in detail here. Since the operation flow is the same as that of the basic embodiment, the operation will be described with reference to the flowchart of FIG.

Now, in steps A1 and A2, the user inputs "I was separated from the tour operator", and in steps A3 and A2
Assume that the following sentence example is searched in A6. Phrase example 4: I got separated from ｛my friend｝.｝ Person｝｝｝ ぐ こ こ こ こ こ こ こ こ こ こ こ こ こ こ こ こ す る す る こ こ こ こ こ こ こ こ. At this time, step A of the basic embodiment
When the degree of coincidence with the input sentence of sentence example 4 is calculated based on No. 8,
Of the semantic features of Sentence Example 4, "person", "separate" and "taita", "separate" and "taita" are in the input sentence. The remaining "person" variable semantic features do not exist in the input sentence. In addition, the sentence example 4 does not include the semantic feature of “passenger” in the input sentence.

Therefore, the degree of coincidence with the input sentence of Example 4 is 100 points obtained by subtracting the number of unmatched numbers 2 multiplied by 50 and the number of unmatched numbers 2 multiplied by 100. .

However, since the score of 100 in sentence example 4 is lower than the threshold value of 200, it is deleted in the next step A13. The present invention is directed to a fixed form sentence for an input sentence in which only a portion corresponding to a variable semantic feature is different and a portion corresponding to a variable semantic feature is registered in a word dictionary. Rather than simply calculating as a mismatch,
The corresponding translation can be displayed.

In the present invention, in step A8 of the basic embodiment, "if a variable semantic feature exists in a fixed sentence,
A character string that does not have a corresponding element in the input sentence and has only one semantic feature that does not match the semantic feature of the fixed phrase in the input sentence and that does not match the semantic feature of the fixed sentence is registered in the word dictionary. If they are, the unmatched variable semantic features in the fixed sentence and the unmatched semantic features in the input sentence are regarded as matched, the degree of matching is calculated, and the character string is added to the word dictionary. Create a translation by replacing with a registered translation. "

Now, in sentence example 4, there is a variable semantic feature "person" that does not match the input sentence. Also, in the input sentence, there is only one character string “contraint” that does not match the semantic feature of the fixed sentence. Furthermore, the word dictionary 19 contains "
And its translations are defined. Therefore, the calculation is performed on the assumption that the “person” and the “passenger” coincide.

Then, sentence example 4 and the input sentence have all the same semantic features. By multiplying the number 3 of matches by 100, the degree of matching with the input sentence of Sentence Example 4 is 300 points.
Sentence example 4 and the matching score of 300 are stored in the working memory 5.

At this time, the replacement unit 12 replaces the variable semantic feature element in the Japanese sentence of the fixed sentence stored in the working memory 5 with “the attendant”, and further replaces the variable semantic feature in the English sentence with “a tourconductor”. Then, the bilingual sentence is stored in the working memory 5 in a form in which the sentence is replaced. 300 points of coincidence in sentence example 4 is threshold value 20
Since the point has exceeded zero, the process proceeds to step A9. Subsequent steps are the same as in the basic embodiment. Step A10 to A15
, The sentence example 4 is displayed on the liquid crystal display 4 as shown in FIG. 13, and the series of translation operations ends.

Next, an embodiment of a machine translation apparatus including the contents of the claim described in claim 2 will be described. In the present embodiment, the following bilingual database in which a connectable character string is added immediately after each semantic feature to the bilingual database of the basic example is used.

[0096] Example 1 ': How much is it up to New York? How much is it to New York ?? Up to the place name / ｛, how much is it? How much is it? got separated from ｛my friend｝. ｛Person｝ /｝ and と も, と も, さ え, and｝ were separated. Here, in ｛｝ following /, a set of character strings that can be connected immediately after the variable semantic feature is registered. I have.

The process of translating an input sentence by the machine translation device according to claim 2 is the same as that of the basic embodiment except for step A8. Therefore, steps B1 to B4 that replace step A8 will be described with reference to the flowchart of FIG. Other steps are the same as those in FIG.

Now, in steps A1 and A2, the user inputs "I missed the train station at the station",
It is assumed that the sentence example 4 ′ is searched from the bilingual database in 3 to A6. Here, it is necessary to search for a partial character string that can be replaced with the semantic feature “friend” from the character string “at the tour operator and at the station”. The following steps B1 to B4 are processes for that.

<Step B1> In step B1, a variable semantic feature exists in the fixed sentence, no corresponding element exists in the input sentence, and only a semantic feature in the input sentence that does not match the semantic feature of the fixed sentence is present. And whether a character string that does not match the semantic feature of the fixed phrase is registered in the word dictionary is checked. If the condition is satisfied, the process proceeds to step B2, and if not, the process proceeds to B4. Now, the fixed sentence 4 ′ has a variable semantic feature “person”, there is no corresponding element in the input sentence, and the word dictionary includes the partial character strings “ station"
Headword exists. These partial character strings are called semantic feature candidate character strings. Since there is a semantic feature candidate character string, the process moves to step B2.

<Step B2> It is checked whether or not a set of character strings connectable immediately after the variable semantic feature includes a character string that appears immediately after the variable semantic feature candidate character string in the fixed phrase, and is included. The degree of coincidence is calculated by regarding a semantic feature candidate character string having a character string as a semantic feature, and the process proceeds to step B3.

Now, the set of character strings 接 続, which can be connected immediately after the variable semantic feature, and と も, さ え, と,｝,
Character string "to" immediately after semantic feature candidate character string "conveyor"
Is included, and the “passenger” is regarded as a semantic feature, and the degree of coincidence is calculated in the same manner as in the basic embodiment. Next, the process proceeds to step B3.

<Step B3> A translated sentence is created by replacing the character string having the semantic feature with a translated word registered in the word dictionary. Now, the variable semantic feature of the English sentence of the fixed phrase 4 ′ is translated into “a tour condu”
The translated text replaced with “ctor” is stored in the working memory 5.

<Step B4> If the calculated degree of coincidence is equal to or greater than a certain threshold value, the flow proceeds to step A9, and if it is less than the threshold value, the flow proceeds to step A13. Here, when the threshold value is calculated as 200, the degree of coincidence becomes 300 and exceeds the threshold value, so that the process proceeds to step A9. As a result, a partial character string that can be replaced with the semantic feature “friend” was searched from the character string “at the tour operator and at the station”, and the translation was replaced. Subsequent steps are the same as in the basic embodiment. In steps A10 to A15, the sentence example 4 'is displayed on the liquid crystal display 4 as shown in FIG. 14, and the series of translation operations ends.

Next, an embodiment of a machine translation apparatus including the contents of claim described in claim 3 will be described. Claim 3
The process of translating an input sentence by the machine translation device described in (1) is the same as that of the basic embodiment except for step A8. Therefore,
Here, steps C1 to C4 that replace step A8
Will be described with reference to the flowchart of FIG. Other steps are the same as those in FIG.

Now, in steps A1 and A2, the user inputs "Do you have plum?"
Suppose that the sentence example 5 is searched from the bilingual database in 6. Here, it is necessary to search for a partial character string that can be replaced with the semantic feature “food” from the character string “sumo”. The following steps C1 to C4 are processes for that.

[0106] It is assumed that such an example is stored in Sentence Example 5.

<Step C1> In step C1, a variable semantic feature exists in the fixed sentence, no corresponding element exists in the input sentence, and only a semantic feature that does not match the semantic feature of the fixed sentence in the input sentence. And whether a character string that does not match the semantic feature of the fixed phrase is registered in the word dictionary is checked. If the condition is satisfied, the process proceeds to step C2, and if not, the process proceeds to C4.

Now, the fixed sentence 5 has a variable semantic feature "tamono", there is no corresponding element in the input sentence, and the word dictionary has a partial character string "sumo" in the word dictionary.
There is a headword "peach". Since there is a semantic feature candidate character string, the process moves to step C2.

<Step C2> In step C2, the longest semantic feature candidate character string is selected, and the flow advances to step C3. Now, the semantic feature candidates are “sumo” and “peach”, and “sumo” is longer.
Is regarded as a semantic feature, and the degree of coincidence is calculated in the same manner as in the basic embodiment. Next, the process proceeds to step C3.

<Step C3> In step C3, a character string having the semantic feature is replaced with a translated word registered in the word dictionary to create a translated sentence. The bilingual sentence in which the variable semantic feature of the English sentence of the fixed sentence 5 is replaced with the translated word “a plum” of “sumo” registered in the machine-capable dictionary is stored in the working memory 5.

<Step C4> In step C4, if the calculated degree of coincidence is equal to or greater than a certain threshold value, step A9
If not, the process proceeds to step A13. Here, when the threshold value is calculated as 200, the degree of coincidence becomes 300 and exceeds the threshold value, so that the process proceeds to step A9. As a result, from the character string "sumo", a partial character string that can be replaced with the semantic feature "food" was searched, and the translation could be replaced. Subsequent steps are the same as in the basic embodiment. In steps A10 to A15, the sentence example 5 is the liquid crystal display 4
15 is displayed as shown in FIG. 15, and a series of translation operations ends.

Next, an embodiment of a machine translation apparatus including the contents of claim described in claim 4 will be described. In the present embodiment, the following bilingual database in which the parts of speech of the respective semantic features are added to the bilingual database of the basic example is used.

[0113] It is assumed that such an example is stored.

Here, the part of speech in the word dictionary of variable semantic features is registered after ':' after the variable semantic features. The process of translating an input sentence by the machine translation device according to claim 4 is the same as in the basic embodiment except for step A8. Therefore, steps D1 to D4 that replace step A8 will be described here with reference to the flowchart of FIG. Other steps are the same as those in FIG.

Now, it is assumed that the user inputs "Please carefully pack in a wooden box" in steps A1 and A2, and that sentence example 6 is retrieved from the bilingual database in steps A3 to A6. Here, it is necessary to search for a partial character string that can be replaced with the semantic feature “case” from the character string “Carefully in the wooden box”. The following steps D1 to D4 are processes for that.

<Step D1> In step D1, a variable semantic feature exists in the fixed sentence, no corresponding element exists in the input sentence, and only a semantic feature in the input sentence that does not match the semantic feature of the fixed sentence is present. And whether a character string that does not match the semantic feature of the fixed phrase is registered in the word dictionary. If the condition is satisfied, the process proceeds to step D2, and if not, the process proceeds to D4.

Now, there is a variable semantic feature “case” in the fixed sentence 6, there is no corresponding element in the input sentence, and in the word dictionary, the partial character string “crate carefully to the wooden box” , There is a headword "Polite". Since there is a semantic feature candidate character string, the process moves to step D2.

<Step D2> At step D2, a semantic feature candidate character string whose part of speech matches the part of speech of the variable semantic feature is selected, and the process proceeds to step D3. Now, the part of speech of the variable semantic feature “case” of the fixed phrase 6 in the bilingual database is a noun, while the candidate of the semantic feature is “wood box” and “attentive”, and the part of speech of “wood box” is the variable semantic feature “ Case"
Therefore, "wooden box" is regarded as a semantic feature, and the degree of coincidence is calculated in the same manner as in the basic embodiment. Next, the process proceeds to step D3.

<Step D3> In step D3, a character string having the semantic feature is replaced with a translated word registered in the word dictionary to create a translated sentence, and the process proceeds to step D4. Now, the bilingual sentence in which the variable semantic feature of the English sentence of the fixed sentence 5 is replaced with the translated word “a wooden box” of “wood box” registered in the machine-capable dictionary is stored in the working memory 5. Next, the process proceeds to step D4.

<Step D4> In step D4, if the calculated degree of coincidence is equal to or more than a certain threshold value, step A9 is executed.
If not, the process proceeds to step A13. Here, when the threshold value is calculated as 200, the degree of coincidence becomes 300 and exceeds the threshold value, so that the process proceeds to step A9. As a result, a substring that can be replaced with the semantic feature “case” was searched from the character string “Carefully in the wooden box”, and the translation was replaced. Subsequent steps are the same as in the basic embodiment. In steps A10 to A15, the sentence example 6 is displayed on the liquid crystal display 4 as shown in FIG. 16, and the series of translation operations ends.

[0121]

<Effect of Claim 1> In the invention according to claim 1, when sentence constituent elements other than the variable semantic features almost match and the variable semantic features do not match, the variable from the word dictionary is used. A bilingual translation of a character string corresponding to a semantic feature can be extracted, and a sentence close to the sentence input by the user and the bilingual translation can be reliably displayed.

<Effect of Claim 2> In the invention according to claim 2, when the sentence components other than the variable semantic features almost match and the variable semantic features do not match, the variable semantic feature is extracted from the word dictionary. If there is more than one such semantic feature candidate, and the set of character strings that can be connected immediately after the variable semantic feature includes the character string immediately after the semantic feature candidate By selecting a feature candidate, a sentence close to the sentence input by the user and its translation can be displayed reliably.

<Effect of Claim 3> In the invention according to claim 3, when sentence constituent elements other than the variable semantic features almost match and the variable semantic features do not match, the variable semantic feature is determined from the word dictionary. If there is more than one such semantic feature candidate, by selecting the longest semantic feature candidate from the semantic feature candidates, a sentence close to the sentence input by the user and its translation are extracted. It can be displayed reliably.

<Effect of Claim 4> In the invention according to claim 4, when the sentence constituent elements other than the variable semantic features almost match and the variable semantic features do not match, the variable semantic feature is extracted from the word dictionary. If there is more than one such semantic feature candidate, select a semantic feature candidate that has a part of speech that matches the variable semantic feature part-of-speech to obtain a sentence close to the sentence input by the user. And its translation can be displayed reliably.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of a conversational sentence translator according to the present invention.

FIG. 2 is an external view of a conversation sentence translator according to an embodiment of the present invention.

FIG. 3 is a flowchart of an embodiment of a translation process according to the present invention.

FIG. 4 is an example of a display screen in one embodiment of the translation processing of the present invention.

FIG. 5 is an example of a display screen in one embodiment of the translation processing of the present invention.

FIG. 6 is an example of a display screen in one embodiment of the translation processing of the present invention.

FIG. 7 is an example of a display screen in one embodiment of the translation processing of the present invention.

FIG. 8 is an example of a display screen in one embodiment of the translation processing of the present invention.

FIG. 9 is an example of a display screen in one embodiment of the translation processing of the present invention.

FIG. 10 is a flowchart of an embodiment of a translation process according to the present invention.

FIG. 11 is a flowchart of an embodiment of a translation process according to the present invention.

FIG. 12 is a flowchart of an embodiment of a translation process according to the present invention.

FIG. 13 is an example of a display screen in one embodiment of the translation processing of the present invention.

FIG. 14 is an example of a display screen in one embodiment of the translation processing of the present invention.

FIG. 15 is an example of a display screen in one embodiment of the translation processing of the present invention.

FIG. 16 is an example of a display screen in one embodiment of the translation processing of the present invention.

[Explanation of symbols]

 Reference Signs List 1 CPU 2 liquid crystal display 3 display driver 4 input recognition unit 5 working memory 6 ROM 7 bilingual database 8 semantic feature dictionary 9 variable semantic feature dictionary 10 analysis unit 11 search unit 12 replacement unit 13 control unit 14 liquid crystal display 15 dedicated pen 16 Handwriting recognition board 17 interpreter icon 18 bus 19 word dictionary

Claims (4)

[Claims] 1. A bilingual database in which an example sentence of a conversation in a first language, a sentence in a second language that is a translation thereof, and a semantic feature are registered as a pair, and extracted as a semantic feature and its semantic feature. A semantic feature dictionary that defines and stores a notation to be written, a variable semantic feature dictionary that has the same semantic feature and registers replaceable words, and an input unit that inputs an input sentence for searching the bilingual database A search unit that searches the bilingual database, the semantic feature dictionary and the variable semantic feature dictionary based on an input sentence input from the input unit; and the search unit searches the bilingual database and the semantic feature dictionary. A replacement unit that performs replacement based on the semantic variable semantic feature dictionary for the example sentence that is searched based on the example sentence extracted by the search unit and the replacement unit; Analysis means for calculating a degree of coincidence with an input sentence input from the input means; output means for outputting a translation of an example sentence based on the analysis means; A word dictionary that can be searched for.In the analyzing means, in the input sentence, the semantic feature is not registered in the semantic feature dictionary and the variable semantic feature dictionary, If there is a character string registered in the word dictionary, the degree of matching is determined assuming that the character string matches the variable semantic feature, and the replacement unit converts the character string into a translated word registered in the word dictionary. A machine translator characterized by converting. 2. In the bilingual database, a connectable character string is added after a character string having each semantic feature, and the analyzing means includes, in the input sentence, the semantic feature dictionary and a variable meaning If there is a character string that is not registered in the feature dictionary but is registered in the word dictionary, a character string that can be connected immediately after the semantic feature dictionary and a character string that immediately follows the character string If they match,
2. The machine according to claim 1, wherein the character string matches the variable semantic feature to determine a degree of coincidence, and the replacing unit converts the character string into a translation registered in the word dictionary. Translator. 3. The analysis means, wherein, in the input sentence, the semantic feature is not registered in the semantic feature dictionary and the variable semantic feature dictionary, but the most significant of the character strings registered in the word dictionary. 2. The method according to claim 1, wherein a matching degree is determined for the long character string as a match with the variable semantic feature, and the replacing means converts the character string into a translation word registered in the word dictionary. Machine translation equipment. 4. In the word dictionary, in addition to the first language word and its translation in the second language, the part of speech of the first language word is registered, and the part of speech of each semantic feature is registered in the bilingual database. In the analyzing means, in the input sentence, although the part of speech is not registered as the semantic feature in the semantic feature dictionary and the variable semantic feature dictionary, of the character strings registered in the word dictionary, A character string that matches the part of speech of the feature is determined as a character string that matches the variable semantic feature to determine the degree of matching, and the replacement unit converts the character string into a translated word registered in the word dictionary. The machine translation device according to claim 1.