Indicating and correcting errors in machine translation systems

• US 9,772,998 B2
• Filed: 05/22/2014
• Issued: 09/26/2017
• Est. Priority Date: 03/22/2007
• Status: Expired due to Fees

• Current Assignee: ABBYY Production LLC (ABBYY Software)
• Original Assignee: ABBYY Production LLC (ABBYY Software)
• Inventors: Anisimovich, Konstantin, Danielyan, Tatiana, Selegey, Vladimir, Zuev, Konstantin
• Primary Examiner(s): Lerner, Martin
• Application Number:
  US14/284,833
• Publication Number:
  US 20140257786A1
• Time in Patent Office: 1,223 Days
• Field of Search: 704 2, 704 3, 704 4, 704 9, 704277, 715264, 715265
• CPC Class Codes: G06F 40/51 Translation evaluation
• CPC Class Codes: G06F 40/58 Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation

Abstract

The preferred embodiments provide an automated machine translation from one language to another. The source language may contain expressions or words that are not readily handled by the translation system. Such problematic words or word combinations may, for example, include the words not found in the dictionary of the translation system, as well as text fragments corresponding to structures with low ratings. To improve translation quality, such potentially erroneous words or questionable word combinations are identified by the translation system and displayed to a user by distinctive display styles in the display of a document in the source language and in its translation to a target language. A user is provided with a capability to correct erroneous or questionable words so as to improve the quality of translation.

Claims

1. A method comprising:

electronically translating, by at least one processor during an automated translation process, a first plurality of words in a source language so as to obtain a second plurality of words in a target language, wherein electronically translating comprises;

performing lexical-morphological analysis of the first plurality of words to generate a lexical-morphological structure of at least one sentence in the first plurality of words,

performing syntactic analysis using the lexical-morphological structure of the at least one sentence to generate a language-independent semantic structure,

performing syntactic synthesis based on the language-independent semantic structure to generate the second plurality of words;

identifying first one or more likely erroneous words in the first plurality of words and corresponding second one or more likely erroneous words in the second plurality of words;

displaying, on a display device, the first plurality of words in the source language;

displaying, on the display device, the second plurality of words in the target language;

automatically indicating, on the display device as part of the automated translation process, the first one or more likely erroneous words within the displayed first plurality of words in the source language;

automatically indicating, on the display device as part of the automated translation process, the second one or more likely erroneous words within the displayed second plurality of words in the target language;

receiving a change to the first one or more likely erroneous words; and

modifying the second plurality of words to provide another translation in the target language based on the change in the first one or more likely erroneous words.

2. The method of claim 1, wherein the first one or more likely erroneous words comprise a word with a lexical error.

3. The method of claim 1, wherein the first one or more likely erroneous words comprise a plurality of words with a syntactical error.

4. The method of claim 1, further comprising displaying two adjacent windows comprising one window for words in the source language and another window for translated words in the target language.

5. The method of claim 1, wherein receiving the change comprises displaying alternatives for substitution with the first one or more likely erroneous words that resemble the first one or more likely erroneous words and providing a capability to select at least one of the alternatives.

6. The method of claim 5, further comprising providing a capability to specify a subject of the first plurality of words and adjusting the alternatives to be consistent with the subject.

7. The method of claim 1, wherein automatically indicating the first one or more likely erroneous words comprises indicating different lexical errors or syntactical errors associated with the first one or more likely erroneous words by different distinctive display styles.

8. The method of claim 7, wherein the lexical errors are identified differently from the syntactical errors.

9. The method of claim 1, wherein identifying the first one or more likely erroneous words comprises identifying one or more words that are not stored in connection with electronically translating the first plurality of words.

10. The method of claim 1, wherein electronically translating comprises computing ratings for syntactic constructions, and wherein identifying the first one or more likely erroneous words comprises identifying one or more words corresponding to the syntactic constructions that have low ones of the ratings.

11. A system comprising:

a display device; and

at least one processor configured to;

electronically translate, during an automated translation process, a first plurality of words in a source language so as to obtain a second plurality of words in a target language, wherein, to electronically translate, the processor is to;

perform lexical-morphological analysis of the first plurality of words to generate a lexical-morphological structure of at least one sentence in the first plurality of words,

perform syntactic analysis using the lexical-morphological structure of the at least one sentence to generate a language-independent semantic structure,

perform syntactic synthesis based on the language-independent semantic structure to generate the second plurality of words, and

identify first one or more potentially erroneous words in the first plurality of words and corresponding second one or more potentially erroneous words in the second plurality of words;

display, on the display device, the first plurality of words in the source language;

display, on the display device, the second plurality of words in the target language;

automatically indicate, on the display device as part of the automated translation process, the first one or more potentially erroneous words within the displayed first plurality of words in the source language;

automatically indicate, on the display device as part of the automated translation process, the second one or more potentially erroneous words within the displayed second plurality of words in the target language;

receive a change to the first one or more potentially erroneous words; and

modify the second plurality of words to provide another translation in the target language based on the change in the first one or more potentially erroneous words.

12. The system of claim 11, wherein the first one or more potentially erroneous words comprise a word with a lexical error.

13. The system of claim 11, wherein the first one or more potentially erroneous words comprise a plurality of words with a syntactical error.

14. The system of claim 11, wherein, to receive the change, the processor is further to display alternatives for substitution with the first one or more potentially erroneous words that resemble the first one or more potentially erroneous words and to provide a capability to select at least one of the alternatives.

15. A non-transitory computer-readable medium having instructions stored therein that, when executed by at least one processor, cause the processor to:

electronically translate, by the processor during an automated translation process, a first plurality of words in a source language so as to obtain a second plurality of words in a target language, wherein, to electronically translate, the processor is to;

perform lexical-morphological analysis of the first plurality of words to generate a lexical-morphological structure of at least one sentence in the first plurality of words,

perform syntactic analysis using the lexical-morphological structure of the at least one sentence to generate a language-independent semantic structure,

perform syntactic synthesis based on the language-independent semantic structure to generate the second plurality of words, and

identify first one or more likely erroneous words in the first plurality of words and corresponding second one or more likely erroneous words in the second plurality of words;

display, on a display device, the first plurality of words in the source language;

display, on the display device, the second plurality of words in the target language;

automatically indicate, on the display device as part of the automated translation process, the first one or more likely erroneous words within the displayed first plurality of words in the source language;

automatically indicate, on the display device as part of the automated translation process, the second one or more likely erroneous words within the displayed second plurality of words in the target language;

receive a change to the first one or more likely erroneous words; and

modify the second plurality of words to provide another translation in the target language based on the change in the first one or more likely erroneous words.

16. The computer-readable medium of claim 15, wherein, to receive the change, the processor is further to display alternatives for substitution with the first one or more likely erroneous words that resemble the first one or more likely erroneous words and to provide a capability to select at least one of the alternatives.

17. The computer-readable medium of claim 16, wherein the instructions further cause the processor to provide a capability to specify a subject of the first plurality of words and to adjust the alternatives to be consistent with the subject.

18. The computer-readable medium of claim 15, wherein, to automatically indicate the first one or more likely erroneous words, the processor is further to indicate different lexical errors or syntactical errors associated with the first one or more likely erroneous words by different distinctive display styles.

19. The computer-readable medium of claim 15, wherein, to electronically translate, the processor is further to compute ratings for syntactic constructions, and wherein, to identify the first one or more likely erroneous words, the processor is further to identify one or more words corresponding to the syntactic constructions that have low ones of the ratings.