Apparatus, system, method, and computer program product for resolving ambiguities in translations

• US 8,214,197 B2
• Filed: 09/11/2007
• Issued: 07/03/2012
• Est. Priority Date: 09/26/2006
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation), Toshiba Digital Solutions Corporation (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Kamatani, Satoshi, Chino, Tetsuro, Furihata, Kentaro
• Primary Examiner(s): He, Jialong
• Application Number:
  US11/898,298
• Publication Number:
  US 20080077392A1
• Time in Patent Office: 1,757 Days
• Field of Search: 704 1- 10, 704/277
• US Class Current: 704/2
• CPC Class Codes: G06F 40/45 Example-based machine translation; Alignment
• CPC Class Codes: G10L 15/005 Language recognition

Abstract

A machine translation apparatus includes a translating unit that translates a source language document into a translated document described in a target language, and creates an ambiguous portion that is a word or a sentence having an ambiguity occurred during translation; a storing unit that stores the translated document and the ambiguous portion; a speech receiving unit that receives a speech in the source language; a recognition unit that recognizes the speech received and creates a source language speech sentence as a recognition result; a translation updating unit that updates the translated language document by retranslating a part of a speech content of the source language speech sentence to which the ambiguous portion corresponds, when the source language document includes the ambiguous portion; and a display control unit that displays the updated translated document on a display unit.

Claims

1. A machine translation apparatus comprising:

a document receiving unit configured to receive an input of a source language document described in a source language;

a first translating unit configured to translate the source language document into a translated document described in a target language in a first translation process, and to identify a document ambiguous portion that is a word or a sentence, for which a plurality of candidate translations occurs during the first translation process;

a storing unit configured to store the translated document and the document ambiguous portion;

a speech receiving unit configured to receive a speech in the source language;

a recognition unit configured to recognize the speech received by the speech receiving unit and to create text of a source language speech sentence as a recognition result;

a second translating unit configured to translate the text of the source language speech sentence into the target language;

an extracting unit configured to extract a source language document sentence relating to the source language speech sentence from the source language document;

an updating unit configured to;

select, when the extracted source language document sentence includes the document ambiguous portion, a translated portion of the source language speech sentence corresponding to the document ambiguous portion of the extracted source language document sentence, the translated portion of the source language speech sentence being obtained by a type of ambiguity identification that is the same type of ambiguity identification as was used in the first translation process that identified the document ambiguous portion;

retranslate the source language document using the selected translated portion of the source language speech sentence, and the extracted source language document sentence relating to the source language speech sentence; and

update the translated document stored in the storing unit using the retranslated source language document; and

a display control unit configured to display the updated translated document on a display unit.

2. The apparatus according to claim 1, wherein the extracting unit:

divides the source language document into source language document sentences;

associates a word included in the each source language document sentence with a word included in the source language speech sentence;

calculates each first similarity between the word included in the source language document sentences and the word included in the source language speech sentence; and

extracts one of the source language document sentences that includes a word having a maximum first similarity to the word included in the source language speech sentence.

3. The apparatus according to claim 2, wherein the extracting unit:

calculates, with respect to each range of the source language document that includes at least one of a page, a chapter, a clause, and a paragraph, each second similarity between each of the source language document sentences contained in the range and the source language speech sentence;

retrieves a range that includes a source language document sentence having a maximum second similarity to the source language speech sentence; and

extracts the source language document sentence from the retrieved range.

4. The apparatus according to claim 1, wherein the extracting unit retrieves information on a range of the source language document that includes at least one of a page, a chapter, a clause, and a paragraph from the source language speech sentence, and extracts the source language document sentence from the range of the source language document.

5. The apparatus according to claim 1, wherein

the first translating unit creates the document ambiguous portion for which a plurality of candidate translations occur during selecting a translated word in the first translation process, and

the updating unit, when the extracted source language document sentence includes an ambiguous portion, selects a translated word selected by the second translating unit, as a translated word of the ambiguous portion, the updating unit retranslating the source language document and updating the translated document stored in the storing unit using the retranslated source language document.

6. The apparatus according to claim 1, wherein

the first translating unit creates the document ambiguous portion for which a plurality of candidate translations occur during selecting a dependency of a word in the first translation process, and

the updating unit, when the extracted source language document sentence includes the document an ambiguous portion selects a dependency between words of the source language speech sentence corresponding to the document ambiguous portion of the extracted source language sentence as a dependency of the document ambiguous portion, the updating unit retranslating the source language document and updating the translated document stored in the storing unit using the retranslated source language document.

7. The apparatus according to claim 1, wherein

the display control unit displays, in association with one another, the translated document and information indicating the ambiguous portion for which the plurality of candidate translations occurs.

8. The apparatus according to claim 7, wherein

the storing unit stores the translated document, the ambiguous portion, and type of ambiguity identification that was used in the first translation process that created the document ambiguous portion, and

the display control unit further displays, in association with one another, the ambiguous portion of the translated document and the type of ambiguity identification.

9. A machine translation system comprising:

a display apparatus that displays a source language document described in a source language; and

a machine translation apparatus that translates the source language document into a translated document described in a target language as a translation result, and that is connected to the display apparatus via a network, wherein

the machine translation apparatus includes;

a document receiving unit that receives an input of the source language document;

a first translating unit that translates the source language document into the translated document in a first translation process and identifies a document ambiguous portion that is a word or a sentence, for which a plurality of candidate translations occurs during the first translation process;

a storing unit that stores the translated document and the document ambiguous portion;

a speech receiving unit that receives a speech in the source language;

a recognition unit that recognizes the speech received by the speech receiving unit and creates text of a source language speech sentence as a recognition result;

a second translating unit that translates the text of the source language speech sentence into a translated sentence described in the target language;

an extracting unit that extracts a source language document sentence relating to the source language speech sentence from the source language document;

an updating unit that, when the extracted source language document sentence includes the document ambiguous portion;

selects a translated portion of the source language speech sentence corresponding to the document ambiguous portion of the extracted source language document sentence, the translated portion of the source language speech sentence being obtained by a type of ambiguity identification that is the same type of ambiguity identification as was used in the first translation process that identified the document ambiguous portion;

retranslates the source language document using the selected translated portion of the source language speech sentence, and the extracted source language document sentence relating to the source language speech sentence; and

updates the translated document stored in the storing unit using the retranslated source language document;

a first display control unit that displays the updated translated document on a first display unit; and

a transmitting unit that transmits the document ambiguous portion stored in the storing unit to the display apparatus, wherein

the display apparatus includes;

a receiving unit that receives the document ambiguous portion from the machine translation apparatus; and

a second display control unit that displays the source language document on a second display unit, the source language document including information for indicating occurrence of the document ambiguous portion where the ambiguity occurs based on the document ambiguous portion received by the receiving unit.

10. A machine translation method comprising:

receiving, by a document receiving unit configured to receive, an input of a source language document described in a source language;

first-translating, by a first translating unit configured to translate, the source language document into a translated document described in a target language in a first translation process, and identifying a document an ambiguous portion that is a word or a sentence, for which a plurality of candidate translations occurs during the first translation process;

storing, by a storing unit configured to store, the translated document and the document ambiguous portion in the storing unit;

receiving, by a speech receiving unit configured to receive, a speech in the source language;

recognizing, by a recognition unit configured to recognize, the speech and creating text of a source language speech sentence as a recognition result;

second-translating, by a second translating unit configured to translate, the text of the source language speech sentence into a translated document sentence described in the target language;

extracting, by an extracting unit configured to extract, a source language document sentence relating to the source language speech sentence from the source language document;

selecting a translated portion of the source language speech sentence corresponding to the document ambiguous portion of the extracted source language document sentence, the translated portion of the source language speech sentence being obtained by a type of ambiguity identification that is the same as the type of ambiguity identification as was used in first translation process that identified the document ambiguous portion;

retranslating the source language document using the selected translated portion of the source language speech sentence, and the extracted source language document sentence relating to the source language speech sentence;

updating the translated document stored in the storing unit using the retranslated source language document; and

displaying, by a display control unit configured to display, the updated translated document on a display unit.

11. A computer program product having a non-transitory computer readable medium including programmed instructions for machine translation, wherein the instructions, when executed by a computer, cause the computer to perform:

receiving an input of a source language document described in a source language;

first-translating the source language document into a translated document described in a target language in a first translation process, and identifying a document ambiguous portion that is a word or a sentence, for which a plurality of candidate translations occurs during the first translation process;

storing the translated document and the document ambiguous portion in a storing unit;

receiving a speech in the source language;

recognizing the speech and creating text of a source language speech sentence as a recognition result;

second-translating the text of the source language speech sentence into a translated sentence described in the target language;

extracting a source language document sentence relating to the source language speech sentence from the source language document;

selecting a translated portion of the source language speech sentence corresponding to the document ambiguous portion of the extracted source language document sentence, the translated portion of the source language speech sentence being obtained by a type of ambiguity identification that is the same type of ambiguity identification as was used in the first translation process that identified the document ambiguous portion;

retranslating the source language document using the selected translated portion of the source language speech sentence, and the extracted source language document sentence relating to the source language speech sentence;

updating the translated document stored in the storing unit using the obtained retranslation of the source language document; and

displaying the retranslated document on a display unit.