JP2977087B2 - Machine translation equipment - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a machine translation for performing mutual conversion between natural languages, and more particularly, to a machine translation device capable of determining a priority order of an editing operation.

2. Description of the Related Art Many conventional machine translation system devices require various editing operations (hereinafter referred to as pre-editing) for a plurality (hereinafter referred to as manuscripts) expressed in a certain language. In that case, there were the following two systems in the prior art. For each sentence: 1. The user determines whether pre-editing is necessary, and if so, the user performs the editing work.

2. The system determines whether or not pre-editing is necessary, and if so, the user performs the editing work.

Further, when a manuscript is composed of a large number of sentences, the following disadvantages are raised with respect to the above two methods. That is, the burden of the user's labor is large with respect to the first method. In the above two methods, it often takes a lot of time for the system to determine whether to perform pre-editing, which leads to a reduction in the efficiency of the entire translation process.

As a known document relating to the first method, for example, there is JP-A-62-203270, which discloses a document to be translated on a display unit and presents it to an operator.
It has an instruction unit for designating a range of a translation target document to be determined as an insertion phrase. As a publicly known document relating to the above two methods, there is, for example, Japanese Patent Application Laid-Open No. 62-203266. Is something you can do.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and a first object of the present invention is to automatically determine whether to perform preprocessing and reduce the burden on the user.
In order to determine whether or not to perform pre-editing, a process that does not require much time is proposed, so that the processing efficiency of the entire translation is not reduced. The second purpose is not only to determine which sentence should be edited, but also to prioritize editing work from which sentence to edit,
The purpose is to allow the amount of editing work to be adjusted according to the translation quality desired by the user.

In order to achieve the above object, the present invention provides an input device for inputting a plurality of sentences expressed in one language, an editing unit for performing various editing operations on the plurality of sentences, Using the result as an input, using the grammar rules and dictionary information and the like, translating the result into one expressed in the other language, and outputting the contents, and a storage for storing the grammar rules and dictionary information And a machine translation device that includes at least a display device for displaying the translation result, and translates from one language to another language.
It is characterized in that a plurality of sentences are divided for each sentence, the length of the sentence is calculated, and instructions for performing editing work are displayed in order from the sentence with the longest sentence length. Hereinafter, a description will be given based on examples of the present invention.

FIG. 1 is a block diagram for explaining an embodiment of a machine translation apparatus according to the present invention, wherein 1 is an input device, 2 is an editing unit,
Reference numeral 3 denotes a translation unit, 4 denotes a storage unit, and 5 denotes a display device. According to the present invention, in the above two methods, whether or not to pre-edit each sentence is determined based on the length of the sentence. The grounds that can be used as a basis for the judgment are as follows.

1. At the technical level of conventional machine translation, long sentences have a lower success rate than short sentences. Even if successful, the quality is low.

2. At the technical level of conventional machine translation, analysis is mainly performed for each sentence, and processing between sentences is hardly performed. Therefore, there is no necessity to edit the document in the order in which it appears.
That is, the present invention provides an input device 1 for inputting a plurality of sentences (documents) expressed in one language A, and an editing unit 2 for performing various editing operations (pre-editing) on the plurality of sentences (documents). A translation unit 3 which receives the result of the editing unit 2 as an input, translates the result into a language expressed by the other language B using grammar rules and dictionary information, and outputs the contents thereof; A machine translation device that includes at least a storage unit 4 for storing information and a display device 5 for displaying a translation result of the translation unit, and that translates one language A into another language B. Further, the translator divides a plurality of sentences (original) into sentences, calculates the length of the sentences (the number of characters or words constituting each sentence), and pre-edits the sentences in order from the longest sentence. It is configured to display an instruction to perform.

FIG. 2 is a flowchart for explaining the operation of the machine translation apparatus of the present invention. When the start is started, a manuscript is input from the input unit (Step 1). The translator divides the input manuscript into sentences and calculates the number of characters in the sentence (St
ep2). If necessary, in step 3, the dictionary information of the storage unit 4 is used for each sentence to extract various information necessary for later translation.

The above information is stored in a table as shown in FIG. 3 (Step 4). The table shown in FIG. 3 is a table in which the serial number of the sentence, the sentence length, and other information are entered. The serial number of the sentence is assigned in the order of appearance in the document for each sentence. Next, if necessary, a word (unknown word) not registered in the dictionary is presented to the user, and a process for prompting registration of dictionary information is performed (Step 5). Next, the user is asked to determine whether or not to perform pre-editing (Step 6). If not, the process proceeds to translation processing (Step 9). When performing, based on the information in the table, from among the sentences that have not been pre-edited, present the sentences to be edited on the display unit as shown in FIG. Encourage editing work for the sentence (Step 7). When editing the sentence is completed, the user is prompted to determine whether to continue the pre-editing (Step
8) If not, perform the next translation process (Step
9). To continue, repeat the above procedure and end.

Note that the display method in FIG. 4 is not limited to the illustrated example.
In the above example, the number of characters is considered as the sentence length. Alternatively, the sentence may be divided into words registered in the dictionary and the number of words may be used.

Effect As described above, according to the present invention, when a manuscript is composed of a large number of sentences, it is automatically determined which sentence is to be pre-edited, so that the burden on the user is small. In addition, since the processing for the above determination can be performed in a relatively simple and short-time processing of counting the number of characters, the efficiency of the entire translation processing is not reduced. The priority of editing work is also determined and presented, so if you want to obtain a result in a short time even with low-quality translation, improve the translation quality per unit editing time by interrupting the editing work and translating. You can adjust the amount of editing work while maximizing the degree.

[Brief description of the drawings]

FIG. 1 is a block diagram for explaining an embodiment of the machine translation apparatus of the present invention, FIG. 2 is a flowchart for explaining the operation of the machine translation apparatus of the present invention, and FIG. FIG. 4 is a diagram showing a table in the flowchart of FIG. 4. FIG. 4 is a diagram showing sentences to be edited in descending order of sentence length from among sentences that have not been pre-edited in the flowchart of FIG. . 1 input device, 2 editing unit, 3 translation unit, 4
Storage unit 5, display device.

Claims (1)

(57) [Claims] 1. An input device for inputting a plurality of sentences expressed in one language, an editing unit for performing various editing operations on the plurality of sentences, a result of the editing unit being input, and a grammar Using a rule and dictionary information and the like, translating into a language expressed in the other language, and outputting the contents thereof; and a translation device including a storage unit for storing the grammar rules and dictionary information. In a machine translation device having at least a display device for displaying a translation result and translating from one language to another language, a plurality of sentences are divided for each sentence, a sentence length is calculated, and a sentence length is calculated. A machine translation device for displaying instructions for performing editing work in order from a long sentence.