Machine translation system having idiom processing function

• US 5,644,774 A
• Filed: 04/25/1995
• Issued: 07/01/1997
• Est. Priority Date: 04/27/1994
• Status: Expired due to Term

• Current Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.), Wisconsin Alumni Research Foundation (University of Wisconsin System)
• Original Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Inventors: Kutsumi, Takeshi, Sata, Ichiko, Okunishi, Toshiyuki, Fukumochi, Yoji
• Primary Examiner(s): Hofsass, Jeffery
• Assistant Examiner(s): HILL, ANDREW M
• Application Number:
  US08/428,547
• Time in Patent Office: 798 Days
• Field of Search: 395/752, 395/754, 395/755, 395/757, 395/759, 395/760, 395/794, 395/710
• US Class Current: 704/4
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking
• CPC Class Codes: G06F 40/55 Rule-based translation

Abstract

A machine translation system having an idiom processing function is disclosed, which includes a keyboard for inputting a word sequence of a first language; a dictionary memory for storing therein idioms of the first language including at least two fixed portions and a variable portion interposed therebetween as headers and translated expressions in a second language corresponding to the respective headers; and a control processor for performing a registration process for newly or additionally registering a header of the first language and translated expressions in the second language corresponding to the header into the dictionary memory, a dictionary lookup process for retrieving a header corresponding to an idiom included in the word sequence of the first language input by means of the keyboard from the headers stored in the dictionary memory by comparing the word sequence of the first language with each of the headers, and an idiom processing process for normalizing an arrangement of fixed portions in a word sequence of the first language which is identified with one of the headers in the dictionary lookup process.

Claims

1. A machine translation system having an idiom processing function, comprising:

an input means for inputting a word sequence of a first language;

a dictionary means for storing therein idioms of the first language including at least two fixed portions and a variable portion interposed therebetween as headers and translated expressions in a second language corresponding to the respective headers;

a registration means for newly or additionally registering a header of the first language and translated expressions in the second language corresponding to the header into said dictionary means;

a dictionary lookup means for retrieving a header corresponding to an idiom included in the word sequence of the first language input by said input means from the headers stored in said dictionary means by comparing the word sequence of the first language with each of the headers; and

an idiom processing means for normalizing an arrangement of fixed portions in a word sequence of the first language which is identified with one of the headers by said dictionary lookup means.

2. A machine translation system as set forth in claim 1,

wherein said dictionary means stores therein the idioms of the first language in such a manner that a principal fixed portion of each of the idioms can be distinguished from a supplementary fixed portion thereof, and

said dictionary lookup means generates part-of-speech information for each word and syntactical information including type information assigned to the fixed portions in the word sequence of the idiom identified with one of the headers for distinguishing the principal fixed portion from the supplementary fixed portion and pointer information indicative of interrelation between words in the input word sequence.

3. A machine translation system as set forth in claim 2,

wherein said idiom processing means prepares information for presuming that a word corresponding to the supplementary fixed portion selected from words identified with the fixed portions of the header of the idiom has been moved to a position of a word corresponding to the principal fixed portion within the input word sequence, and adds the information to a word other than the word corresponding to the supplementary fixed portion to transform the syntactical information when carrying out the normalization.

4. A machine translation system as set forth in claim 2,

wherein said idiom processing means prepares information for presuming that a word corresponding to the supplementary fixed portion selected from words identified with the fixed portions of the header of the idiom has been deleted from the input word sequence, and adds the information to a word other than the word corresponding to the supplementary fixed portion to transform the syntactical information when carrying out the normalization.

5. A machine translation system as set forth in claim 2,

wherein said idiom processing means prepares information for presuming that words identified with the variable portions of the header of the idiom has been moved to different positions within the input word sequence, and adds the information to a word other than the words identified with the variable portions to transform the syntactical information when carrying out the normalization.

6. A machine translation system as set forth in claim 2,

wherein the idioms of the first language stored in said dictionary means include a split idiom having at least two fixed portion and a variable portion interposed therebetween.

7. A machine translation system as set forth in claim 6,

wherein, when a split idiom having first and second fixed portions and variable portion interposed therebetween is subjected to the normalization, said idiom processing means prepares information for presuming that a word corresponding to the supplementary fixed portion selected from the first and second fixed portions has been deleted, and adds the information to a word other than the word corresponding to the supplementary portion to transform the syntactic information.

8. A machine translation system as set forth in claim 2, further comprising:

a translation generating means for generating a translation from said syntactical information normalized by said idiom processing means; and

an output means for outputting said translation.