JPH05334349A - Machine translating system - Google Patents

Abstract

PURPOSE:To provide the machine translating system capable of easily understanding the contents of an output part in machine translation. CONSTITUTION:When a sentence 'I read the book that he bought yesterday.' is inputted as shown in (A), the output result of a syntax analysis is obtained as shown in (B). Then the syntax tree is deformed as shown in (C). Here, it is requested that 'buy' and 'book' originating from 'bought book' in Japanese form a connection clause. Consequently a relation clause generation rule generates new nodes [RELCL] and [RP], and asterisks as new nodes. A corresponding English sentence is therefore generated as shown in (E) and the part of the relation clause is bracketed and displayed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine translation device,
In particular, the present invention relates to a machine translation device capable of clearly indicating syntax information so that an output language can be efficiently understood.

[0002]

2. Description of the Related Art A machine translation device that is interested in the present invention is disclosed in, for example, Japanese Patent Application Laid-Open No. 2-156377. According to the publication, the dependency relation is displayed in the output language after translation. That is, by the operator performing a specific operation, the result of the dependency analysis of the output language is displayed as a two-dimensional figure in a region different from the region displaying the output language.

[0003]

The conventional machine translation apparatus is configured as described above. However, as in the past, only the result of the dependency analysis was displayed as a two-dimensional figure in a region different from the region for displaying the output language, so it can be referred only when the operator feels doubt about the output language. If the operator misunderstands, the misunderstanding cannot be resolved. Also, displaying as a two-dimensional figure lacks intuition and hinders understanding of the output language.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a machine translation device which allows an operator to intuitively and efficiently understand an output language without misunderstanding. .

[0005]

A machine translation device for converting a sentence in an input language into a sentence in a target language based on a dictionary and grammatical rules according to the present invention linguistically analyzes a sentence in an input language. The language analysis means includes a detecting means for detecting the presence of a portion having a specific syntactic relation in the input sentence, and the language analyzing means is detected when the detecting means detects the portion having the syntactic relation. And a display means for displaying a part having a syntactic relationship and displaying the part together with the converted sentence.

[0006]

If there is a portion having a specific syntactic relation in the sentence of the input language, the portion is specified and translated and then displayed simultaneously in the converted sentence. Since the part having a specific syntactic relation is displayed in the translated sentence, the operator can easily understand the part having the syntactic relation.

[0007]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a main part of a machine translation device according to the present invention. With reference to FIG. 1, a machine translation apparatus includes an input unit 1 that reads an input language to be machine translated.
And a morphological analysis processing unit 2 that divides a sentence input based on the dictionary 7 and the grammatical rule 8 into morpheme units and acquires a part of speech and other information. Not only input as file information but also text may be directly input to the input unit 1 from a keyboard (not shown).

The machine translation device further displays a specific syntactic structure and a syntactic analysis processing unit 3 that analyzes the dependency relations and case relations in the sentences input using the dictionary 7 and the grammar rules 8. A generation processing unit 4 that includes the rules to be performed and that generates an output sentence after translation, a display unit 5 that displays an output language that is a translation result, and a storage unit 6 that stores the translation result.

FIG. 2 is a flow chart showing the flow of processing in the generation processing unit 4 shown in FIG. Referring to FIG. 2, generation processing unit 4 first performs a syntax analysis output process of receiving a syntax tree which is output information from syntax analysis processing unit 3 shown in FIG. 1 (S11). Next, the syntax tree is sequentially transformed (S12). In this case, rules for displaying a specific syntactic structure are applied if necessary. A specific example to which this rule is applied will be described later. Next, it is transformed into a syntax tree of the output language (S13). Next, the end node of the syntax tree of the output language is transformed into an appropriate morpheme, and then arranged one-dimensionally in order to form the output language (S1
3), the result is output to the display unit 5 (S14).

Next, the specific contents of the operation of the generation processing unit 4 shown in FIG. 2 will be described. For example, a machine translation system may be I read the book that that
boot yesterday. When outputting the English sentence, it is unclear which of the following meanings should be understood.

Sentence (1): "I read the book he bought yesterday" Sentence (2): "I read the book he bought yesterday" So according to the invention the above two sentences are It is translated and divided as follows.

3 and 4 are translated from Japanese into English by the machine translation device according to the present invention, and the relative clauses led to relative pronouns and relative adverbs are clearly indicated in the English sentence as "specific syntactic structure". It is a figure explaining the example in the case of doing. Here, first, as shown in FIG. 3A, a case will be described in which the above sentence (1) "I read a book he bought yesterday" was input. First, (B) is a diagram showing output information of syntax analysis (S11 in FIG. 2). When the output process is completed, the Japanese sentence is displayed like this. Next, the syntax tree is transformed as shown in FIG.
Here, the state immediately before the rule for displaying the relative clause is applied is shown. Here, "buy" and "book", which are derived from the Japanese word "buy a book", require the formation of a relative clause.

Here, each symbol has the following meaning.

[0015]

[Table 1] [S] sentence or subordinate clause [NP] noun phrase [AX] auxiliary verb [VP] verb phrase [RLCL] relative clause [RP] relative pronoun [DT] article [AVP] adverb phrase ★ for displaying syntactic structure Node Next, according to the relative clause generation rule, the new node [RLCL]
FIG. 3D shows a state in which the syntax structure display node is generated as a new node at the same time as and [RP] are generated. Then, as shown in FIG. 4A, the final result of the transformation of the syntax tree is obtained. This syntax tree is sent to the sentence generation step shown in S13 of FIG. 2, and the result is output to the display unit 5 as shown in FIG. 3 (E) (FIG. 2, S14).

In this example, [] is used to distinguish the relational clause from other parts, but the display method is not limited to this.

Next, referring to FIGS. 5, 6 and 7, when the sentence (2) shown in FIG. 5A, "I read the book he bought yesterday" is input. An example will be described. Also in this case, the syntactic analysis result is output as shown in FIG.

Next, the syntax tree is transformed as shown in FIG. Here, "h" that comes from the Japanese "book he bought"
e "," buy "and" book "require the formation of a relative clause.

According to the relative clause generation rule, the new node [RE
LCL], [RP], [S], etc. are generated, and at the same time, the syntactic structure display node * is generated as a new node. Then, as shown in FIG. 7A, the final result of the transformation of the syntax tree is obtained. Then, the sentence shown in FIG. 7B is displayed on the display unit 5.

In the above embodiment, the relational pronoun was given as an example of the specific syntactic structure, but the present invention is not limited to this, and it goes without saying that it is applied to sentences of parallel structure and the like.

[0021]

As described above, according to the present invention, in a sentence of an input language, if there is a portion forming a specific syntax, that portion is specified and displayed. Since the part having a specific syntactic relation is specified and displayed together with the translated sentence, the operator can easily understand the part having the syntactic relation. As a result, in machine translation, the operator can easily understand the contents of the output unit.

[Brief description of drawings]

FIG. 1 is a block diagram showing a main part of a machine translation device according to the present invention.

FIG. 2 is a flowchart for explaining the operation of the generation processing unit according to the present invention.

FIG. 3 is a diagram showing a first example of a case where the machine translation device according to the present invention is applied to translation from Japanese to English.

FIG. 4 is a diagram showing a first example when the machine translation device according to the present invention is applied to translation from Japanese to English.

FIG. 5 is a diagram showing a second example in which the machine translation device according to the present invention is applied to translation from Japanese to English.

FIG. 6 is a diagram showing a second example in which the machine translation device according to the present invention is applied to translation from Japanese to English.

FIG. 7 is a diagram showing a second example when the machine translation device according to the present invention is applied to translation from Japanese to English.

[Explanation of symbols]

 1 Input unit 2 Morphological analysis processing unit 3 Syntax analysis processing unit 4 Generation processing unit 5 Display unit

Claims (1)

[Claims] 1. A machine translation device for converting a sentence of an input language into a sentence of a target language based on a dictionary and a grammar rule, including a means for analyzing the sentence of the input language, the language analysis The means includes a detection means for detecting the presence of a portion having a specific syntactic relationship in the input sentence, and has the detected syntactic relationship when the detection means detects the syntax. A machine translation device, comprising: a display unit that identifies a part and displays it in the converted sentence.