COMPUTER-ASSISTED NATURAL LANGUAGE TRANSLATION

• US 20100223047A1
• Filed: 12/14/2009
• Published: 09/02/2010
• Est. Priority Date: 03/02/2009
• Status: Active Grant

• Current Assignee: SDL Ltd.
• Original Assignee: SDL PLC (RWS Holdings Plc)
• Inventors: Christ, Oliver
• Granted Patent: US 8,935,148 B2
• Time in Patent Office: Days
• US Class Current: 704/4
• CPC Class Codes: G06F 40/274 Converting codes to words; Guess-ahead of partial word inputs
• CPC Class Codes: G06F 40/47 Machine-assisted translation, e.g. using translation memory

Abstract

A computer implemented method of translating source material in a source natural language into a target natural language includes receiving a first data input which is a first part of a sub-segment of a translation of the source material from the source natural language into the target natural language, identifying a selectable target text sub-segment in the target natural language associated with the received first data input, and outputting the selectable target text sub-segment. The selectable target text sub-segment is extracted from a corpus of previously translated text segment pairs, each text segment pair having a source text segment in the source natural language and a corresponding translated text segment in the target natural language.

Claims

1. A computer implemented method for use in translation of source material in a source natural language into a target natural language, said method comprising performing, by a processor in a software process, the steps of:

receiving a first data input in said target natural language, said first data input comprising a first part of a sub-segment of a translation of said source material from said source natural language into said target natural language;

identifying at least one selectable target text sub-segment in said target natural language associated with said received first data input, said at least one selectable target text sub-segment having been extracted from a corpus of previously translated text segment pairs, each text segment pair comprising a source text segment in said source natural language and a corresponding translated text segment in said target natural language; and

outputting said at least one selectable target text sub-segment.

2. A method according to claim 1, wherein said identified target text sub-segment is output in a form suitable for review by a translation system operator, whereby said at least one identified target sub-segment can be selected for use in a translation of said source material into said target natural language.

3. A method according to claim 1, comprising receiving a second data input, said second data input comprising a selection of a target text sub-segment from said at least one outputted target sub-segment for use in a translation of said source material into said target natural language.

4. A method according to claim 1, comprising the extraction of target text sub-segments from a corpus of previously translated text segment pairs, wherein extraction of target text sub-segments from said corpus comprises computation of a measure of co-occurrence between words in said source text segments and words in said corresponding translated text segments in said previously translated text segment pairs.

5. A method according to claim 1, wherein a plurality of selectable target text sub-segments are output.

6. A method according to claim 5, wherein the number of selectable target text sub-segments which are output is limited by a predetermined, user-configurable threshold.

7. A method according to claim 5, comprising emphasising in said output a given target text sub-segment, selected from said plurality of selectable target text sub-segments as a best initial selection.

8. A method according to claim 1, wherein said first data input comprises a number of text characters, and said identifying step comprises identifying target text sub-segments having text characters corresponding to text characters in said first data input.

9. A method according to claim 8, wherein said at least one selectable target sub-segment is output for review by a translation system operator in response to said first data input reaching a predetermined number of text characters.

10. A method according to claim 9, wherein said predetermined number of text characters is user-configurable.

11. A method according to claim 1, wherein said identifying step comprises identifying a plurality of target text sub-segments in said target natural language associated with said received first data input, said plurality of target text sub-segments having been extracted from said bilingual corpus, said method comprising:

receiving a third data input in said target natural language, said third data input comprising a second part of a translation of said source material from said source natural language into said target natural language;

generating a subset of selectable target text sub-segments from said plurality of identified target text sub-segments associated with said received first and third data inputs,

wherein said outputting comprises outputting said generated subset of selectable target text sub-segments for review by said translation system operator.

12. A method according to claim 1, wherein in the event that a plurality of target text sub-segments are identified in said identifying step, said method comprises ranking said identified target text sub-segments according to the amount of source material in said source natural language and/or target material in said target natural language represented by each of said identified target text sub-segments in said target natural language,

wherein in said outputting step, said identified target sub-segments are output in order of rank.

13. A method according to claim 1, wherein a target text sub-segment comprises a word or phrase in said target natural language.

14. A method according to claim 1, wherein a text segment pair comprises a sentence or paragraph in said source natural language and a corresponding translated sentence or paragraph in said target natural language.

15. A method according to claim 1, comprising:

identifying in said source material one or more placeable element, a placeable element being an element of source material that does not require translation into said target natural language by a translation system operator and which can be converted from said source natural language into said target natural language using one or more predetermined rules;

converting said one or more identified placeable element into said target natural language using said one or more predetermined rules and inserting said placeable element into said full translation without review or translation by said translation system operator.

16. A method according to claim 15, wherein said identified placeable element comprises one or more of:

a time,

a date,

a numerical expression,

a measurement expression,

an acronym,

a tag,

an item of formatting, and

a name.

17. A method according to claim 1, wherein said corpus comprises one or more of the following:

a translation memory,

one or more lookup results from a terminology database,

an aligned bilingual text, and

a user-definable list of target sub-segments.

20. A computer program product comprising a computer readable medium having computer readable instructions recorded thereon, the computer readable instructions being operative, when performed by a computerised device, to cause the computerised device to perform the method of claim 1.

18. A computer implemented method for use in translation of source material in a source natural language into a target natural language, said method comprising performing, by a processor in a software process, the steps of:

storing a set of target text sub-segments for use in translation of said source material into said target natural language;

providing a user interface, said user interface including a display part for displaying a segment of said source material in the source natural language and a text entry part for inputting textual characters in the target natural language;

receiving, via said text entry part, a first data input in said target natural language, said first data input comprising a first part of a translation of some of said segment of source material from said source natural language into said target natural language;

identifying, on the basis of said received first data input, at least one selectable target text sub-segment from said set of target text sub-segments; and

outputting said at least one selectable target text sub-segment for use in said translation of some of said segment of source material into the target natural language.

19. A natural language translation apparatus for use in translation of source material in a source natural language into a target natural language, said apparatus including a processor configured to:

receive a first part of a sub-segment of a translation of said target material from said source natural language into said target natural language;

identify at least one selectable target text segment in said target natural language associated with said received first data input, said at least one selectable target text sub-segment having been extracted from a corpus of previously translated segment text pairs, each text segment pair comprising a source text segment in said source natural language and a corresponding translated text segment in said target natural language; and

output said at least one selectable target segment.