JPS5884364A - Interpreter - Google Patents

Abstract

PURPOSE:To obtain a correctly interpreted sentence without erroneous interpretation and with high efficiency, by performing the interpretation for each word or each word train which is divided with the punctuated information and the punctuated points of a sentence. CONSTITUTION:A interpretation mode is obtained through an input device 7, and a sentence to be interpreted is fed to the main body of an interpreter 1. This sentence is converted into a pattern code by a controller 3 and applied to a display 5 to be displayed in the form of a train of a character pattern. Then an operator designates and feeds a prescribed display position on a screen via a tablet devide 6 to the original sentence displayed at the display 5. Thus the position information is fetched into the controller 3, and the designated display character pattern blinks. The controller 3 interpretes the character pattern in the form of the punctuated information of the word or word train and then divides the word train within a designated range. Then the device 7 gives a command for interpretation.

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field of the Invention The present invention relates to a highly practical translation device that can efficiently translate between natural languages without impairing the naturalness of the languages.

TECHNICAL BACKGROUND OF THE INVENTION Recently, devices have been attracting attention that analyze the structure of a given sentence or sentence according to programmed grammar rules and translate the sentence or the entire sentence by referring to a mechanical dictionary registered in a storage device. ing.

According to this device, it is possible to improve the efficiency of translation work, which conventionally relied on manual labor, and it is possible to easily automatically print out the translated text. Moreover, it has the advantage that specialized terms and knowledge can be registered in advance in a machine dictionary and then searched.0Problems with the Background ArtHowever, the translation work using such a device is only 70 grams. Since syntactic analysis follows grammatical rules and is fixedly given by a mechanical dictionary cape, it is extremely difficult to make appropriate translations using one-dimensional meaning and context information. As a result, there is a problem in that only a literal translation of iJ1 is obtained, and the meaning of the words cannot be sufficiently reflected. Moreover, in bad cases, semantically clear translations are often made. The purpose of this invention is to provide a highly practical translation device that can efficiently translate between appropriate natural words without impairing the naturalness of the language.

Summary of the Invention The present invention displays a given sentence or text as a string of characters/4' keys on a display, and inputs delimiter information given to the displayed sentence or text in relation to its context, etc. The gist is a translation device that translates the sentence or sentence by executing translation processing for each word or word string inputted through the device and divided by the above-mentioned delimiter) information and the sentence delimiter point. It is something to do.

Therefore, according to the present invention, it becomes possible to perform translation in an interactive manner between the device and the operator, and it becomes possible to perform appropriate translation while reflecting the meaning of the context.

Furthermore, even for ambiguous 6#& (words or word strings), by giving them a break that takes into account the context, it is possible to perform appropriate translations, without sacrificing the naturalness of the language. In addition, the above-mentioned purpose is quickly achieved by demonstrating the tremendous effect of gin, which can efficiently obtain correct translations without causing obvious semantic mistranslations.

Embodiment of the Invention Hereinafter, an embodiment of the present invention will be described with reference to Figure t11.
I will clarify.

No. 11 is a schematic diagram of the embodiment device, J is Tsumugi BR,
This is the main body of the device. This translation device main body 1 includes a storage device 2 that stores a grammar rule program and a machine dictionary given in advance, and a control device 3 that controls the parsing and translation processing of a given sentence or sentences by the main body 1. is connected. Further, an external storage medium input/output device 4 is connected to the translation device main body J via a control device 3.

The input/output device 4 is composed of an external large-capacity memory such as a magnetic disk or a magnetic drum, and stores the mechanical dictionary and the like classified by item, for example.

Therefore, the sentence or text to be subjected to translation processing, for example, an English gambling sentence, found in the translation device main body IK4 is transferred to the control device 3.
The text is supplied to a display device 5 consisting of a CRT display panel, and is displayed as a string of letter turns representing the text. Further, a tablet device 6 is incorporated on the display surface of the display device 5, and is configured to input positional information with respect to the display screen. Further, the input device shown in 1 in the figure is configured to receive various control information related to the above translation, such as from a keymate device, etc.
Control information input via the tablet device 6 or the input device 1'ft is given to the control device 3, analyzed, and used to control the translation process. The translation of the English candy sentence, for example, a Japanese sentence, translated under the control of the control device 30 is supplied to the display device 5, where it is converted into a kana/kanji pattern string and displayed.
Depending on the control mode, the information may be output as a printout or the like via the hard copy device 8.

Note that input of designation of screen position information via the tablet device 6 and input of control information via the input device 7 are performed by an operator on the text displayed on the display device 5. Especially in the evening! and measuring device 6
The position i* information is provided manually as information that separates the displayed text, as will be described later. The mechanism of this tablet device 6 may be a light pen type or a touch 79 panel type. Alternatively, the sentence or text to be translated may be provided via the external self-storage output device 4t.

Therefore, when translating from English to Japanese in the apparatus configured as described above, the translation mode is first given via the input device i. Then, the child who goes to the beginning of the sentence that is involved in translation is included in the translation device body 1.

For example, r The capital saw a b@
aut1ful girl with exp@n5iv
The English sentence e f1*1d grams@s, J is given. In the control device S, this English text is converted into an English letter 9 turn code and displayed on the display device 5, and the display device 5 converts this into an English letter 9 turn code, for example, as shown in FIG. 2(a)K. Display as a column of turns. The display surface of the display device 5 is divided into, for example, an area A for displaying the original text, an area B for displaying partial translation results or message text, and an area C for displaying the translated text. At the end of the translated text displayed in area C, a trailer mark D is displayed to indicate this. However, for the original text displayed on the display device 5, the operator first specifies and inputs the light display position of "Th@" and the display position of "captain" on the display screen through the tablet device 6. The information is taken into the control device 3. In this case, ``Th・“
By specifying, the displayed character no l turn will blink, and "
It is convenient to stop the blinking and change the display color by specifying eaptaln ''', because the input process can be accurately grasped.In this way, when both position information are input, the control device 3 interprets this as word or word string delimiter information and classifies the word string in this specified range from the English sentence.Then, by giving a translation command via the input device 1, the word string is [Th
e captain J is translated by the translation device main body 1. This translation is performed by referring to a machine dictionary, extracting the word's part of speech, meaning, and translation information, and then parsing it using Shirutsuchi parsing grammar rules. and "captain"
Obtain Naru Japan @ as a translation word (translation). This translation is then displayed in screen area B of the display device 5 as shown in FIG. 2(b).
will be displayed. Note that if there are two or more candidate sentences for the word string to be translated in this way, all of these candidate sentences are displayed in the area B. Then, by selecting and specifying display sentences (candidate sentences) in O(Rake) via the tablet device 6, the most appropriate candidate sentence is specified, and the others are removed.

In this way, other parts of the original text are also translated interactively. After that, according to the operator's instructions, from the above @'I'h@" to "gtas+s@s
, ” Assuming that the word string leading to K has one moat and one moat, when that information is input, the translation device main unit 1 parses the syntax to compose one sentence by integrating the partially translated sentences that have already been translated. and its translation process is executed.Then, the translated sentences compiled into one K are displayed in the area BK as shown in Fig. 2 (C).In response, the operator determines that it is mourning, and the
By inputting OKJ information through the input device 1,
When the translation of the original text is completed, the resulting translated text moves to area C and is displayed as shown in FIG. Become.

After that, the above-mentioned instruction input operation and translation process are repeatedly executed, and here Man! This happens when thin interactive translation is performed.

In addition, when the above-mentioned translated sentences are combined into one translated sentence K, the relationship between each translated sentence becomes a problem, and sentences with completely different meanings may be created. It is extremely difficult for computer systems to make judgments about text analysis that involves such semantic understanding and context understanding. Therefore, if such information is given by inputting instructions by the operator, it is possible to prevent semantically obvious translations, which is very effective. Therefore, according to this apparatus, it is possible to perform highly efficient, accurate, and high-quality translation through the above-mentioned interactive processing, and the practical advantage is enormous.

According to the present device discussed above, the following effects can be expected. In other words, the operator is extremely easy to handle, requiring only the position specification on the display screen and the operation of a few command keys, and the efficiency of translation processing can be significantly improved. Since it can be done using a machine dictionary, it is also effective in making it possible to organize translations, such as specialized candy.

Furthermore, since translation processing is executed based on short word strings, accurate translation can be performed with relatively simple processing. Therefore, translation not only between cognate languages but also between different languages is easy with a relatively simple system configuration! It has the following effects: Therefore, its practical advantages are very high.

Note that the present invention is not limited to the above embodiments. For example, it can be applied not only to English-Japanese translations, but also to French-Japanese, German-Japanese, and Japanese-Russian translations.

In short, the present invention can be implemented with various modifications without departing from its essential features.As described in detail, according to the present invention, the present invention can be implemented in an interactive manner, taking into account the meaning and context of the original text. Since the translation is divided into word strings and translated sequentially, and then the translated text is edited, the translation process can be performed simply, accurately, and efficiently, which has a great practical advantage.

[Brief explanation of drawings]

FIG. 1 is a schematic configuration diagram of an embodiment of the present invention, and FIGS. 2(a) to 2(d) show character strings of the original text displayed on the display device and character strings of the translated text. It is a diagram showing the relationship with the turn train. l- Translation device main body, 2... Storage device, 3... Control device, 4... External storage medium input/output device, 5... Display device, 6... Taplet device (position input device), 7.
... Input device (command information), 8... Bird buzzer device. Figure 1 Figure 2 (a) (b) (c) (d)

Claims (1)

[Claims] A display that displays a given dead sentence or sentence as a character j4-turn string, and a display that displays the word or word string delimiter information specified for the sentence or sentence displayed on this display as the character j4-turn string displayed above. A translation characterized by comprising: an input device for specifying and inputting a corresponding position; and means for translating each word or word string divided by the input break information and the break point of the sentence. Device.