Language transference rule producing apparatus, language transferring apparatus method, and program recording medium

• US 7,321,850 B2
• Filed: 01/31/2006
• Issued: 01/22/2008
• Est. Priority Date: 06/04/1998
• Status: Expired due to Fees

• Current Assignee: Panasonic Intellectual Property Corporation of America (Panasonic Holdings Corporation)
• Original Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Inventors: Wakita, Yumi
• Primary Examiner(s): OPSASNICK, MICHAEL N
• Application Number:
  US11/344,027
• Publication Number:
  US 20060129381A1
• Time in Patent Office: 721 Days
• Field of Search: 704/10, 704/8, 704/9, 704/277
• US Class Current: 704/10
• CPC Class Codes: G06F 40/45 Example-based machine translation; Alignment
• CPC Class Codes: G06F 40/55 Rule-based translation

Abstract

A language transferring apparatus is provided. A language rule producing section statistically automatically trains grammatical or semantic restriction rules for a partial word or a word string from a parallel-translation corpus. The rules are described such that a source language partial sentence corresponds to a target language partial sentence. A speech recognizing section performs speech recognition on source language speech using the language rules and outputs a recognition result. A language transferring section transfers a source language sentence into a target language sentence using the same language rules. Even when a portion of an input speech sentence contains an untrained portion or when speech recognition is partly erroneously performed, transference to the target language is surely enabled. Moreover, a phrase dictionary and interphrase rules which are necessary for transference can be automatically produced without requiring much manual assistance.

Claims

1. A language transferring apparatus characterized in that said apparatus comprises:

storing means of storing language rules which are obtained by training grammatical or semantic restriction rules for a word or a word string from a training database including a parallel translation corpus in which a source language sentence that is input in a form of speech or text, and that has undergone a corresponding target language transference, is automatically paired with a target language sentence comprising the source language sentence that has undergone the corresponding language transference;

a speech recognizing section which performs speech recognition on input speech by using the stored language rules, and which outputs a result of the recognition in a form of a sentence that is a target language transference;

a language transferring section which transfers the sentence that is the target language transference, into a transferred sentence that has undergone language transference, by using the same language rules as that used in said speech recognizing section; and

a speech synthesizing section configured to aurally output a translation resultant sentence, based on the transferred sentence,

wherein said speech recognizing section and said language transferring section perform each processing by dealing with one bundle defined by a style-independent phrase as a common language processing unit.

2. A language transferring apparatus according to claim 1, characterized in that said apparatus comprises:

a rule distance calculating section which, for a language rule group which is obtained by, among the language rules, bundling language rules of a same target language sentence as a same category, calculates an acoustic rule distance of the sentence that is a target language transference of language rules contained in the language rule group;

an optimum rule producing section which, in order to enhance a recognition level of speech recognition, optimizes the rule group by merging language rules having a similar calculated distance.

3. A language transferring apparatus according to claim 1, wherein said speech synthesizing section performs speech synthesis on the sentence that has undergone language transference, by using a same language rules as that used in said language transferring section.

4. A language transferring apparatus according to claim 3, characterized in that said apparatus comprises:

a rule distance calculating section which, for a language rule group which is obtained by, among the language rules, bundling language rules of a same target language sentence as a same category, calculates an acoustic rule distance of the sentence that is a target language transference of language rules contained in the language rule group;

an optimum rule producing section which, in order to enhance a recognition level of speech recognition, optimizes the rule group by merging language rules having a similar calculated distance.

5. A language transferring apparatus according to claim 1, characterized in that the language rules are produced by dividing the sentence that is the target language transference, and the transferred sentence into portions in which both sentences form semantic consistency having style-independent phrases, and making rules with separating language rules in the style-independent phrases from language rules between the style-independent phrases.

6. A language transferring apparatus according to claim 5, characterized in that said apparatus comprises:

a rule distance calculating section which, for a language rule group which is obtained by, among the language rules, bundling language rules of a same target language sentence as a same category, calculates an acoustic rule distance of the sentence that is a target language transference of language rules contained in the language rule group;

an optimum rule producing section which, in order to enhance a recognition level of speech recognition, optimizes the rule group by merging language rules having a similar calculated distance.

7. A language transferring apparatus according to claim 5, characterized in that the language rules are produced by making rules on grammatical or semantic rules in the style-independent phrases and concurrent or connection relationships between the style-independent phrases.

8. A language transferring apparatus according to claim 7, characterized in that said apparatus comprises:

a rule distance calculating section which, for a language rule group which is obtained by, among the language rules, bundling language rules of a same target language sentence as a same category, calculates an acoustic rule distance of the sentence that is a target language transference of language rules contained in the language rule group;

an optimum rule producing section which, in order to enhance a recognition level of speech recognition, optimizes the rule group by merging language rules having a similar calculated distance.

9. The language transferring apparatus according to claim 1, wherein the source language sentence is automatically paired with the target language sentence without using any grammatical rule.

10. The language transferring apparatus according to claim 9, wherein partial sentences are automatically extracted for each of the source language sentence and the target language sentence, each of the partial sentences formed with semantic consistency without using any grammatical rule, and

the partial sentences of the source language sentence are compared to the partial sentences of the target language sentence with respect to a whole of a sentence to determine corresponding partial sentences.

11. The language transferring apparatus according to claim 10, wherein the partial sentences are automatically extracted by calculating a frequency of adjacency of words or parts of speech in the source language sentence and the target language sentence, and coupling the words and the parts of speech of a high frequency of adjacency.

12. A language transferring apparatus characterized in that said apparatus comprises:

storing means of storing language rules which are obtained by training grammatical or semantic restriction rules for a word or a word string from a training database including a parallel translation corpus in which a source language sentence that is input in a form of speech or text, and that has undergone a corresponding target language transference, is automatically paired with a target language sentence comprising the source language sentence that has undergone the corresponding language transference;

a speech recognizing section which performs speech recognition on input speech by using the stored language rules, and which outputs a result of the recognition in a form of a sentence that is a target language transference; and

a language transferring section which transfers the sentence that is the target language transference, into a transferred sentence that has undergone language transference, by using the same language rules as that used in said speech recognizing section,

wherein said speech recognizing section and said language transferring section perform each processing by dealing with one bundle defined by a style-independent phrase as a common language processing unit, and

a translation resultant sentence is output to be displayed as text, based on the transferred sentence.