Translating keywords from a source language to a target language

• US 8,484,218 B2
• Filed: 04/21/2011
• Issued: 07/09/2013
• Est. Priority Date: 04/21/2011
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Raghunath, Mandayam Thondanur
• Primary Examiner(s): SYED, FARHAN M
• Application Number:
  US13/091,272
• Publication Number:
  US 20120271828A1
• Time in Patent Office: 810 Days
• US Class Current: 707/739
• CPC Class Codes: G06F 16/3337 Translation of the query language, e.g. Chinese to English
• CPC Class Codes: G06F 40/44 Statistical methods, e.g. probability models

Abstract

In one implementation, a method includes receiving a request for translation of one or more first keywords from a source language to a target language; and translating, using a machine translation process, the first keywords from the source language into a plurality of second keywords in the target language. The method can also include determining, by a computer system, frequencies with which each of the second keywords occur in a corpus associated with the target language. The method can further include selecting, by the computer system, a subset of the second keywords to use in the target language based on the determined frequencies of occurrence.

Claims

1. A computer-implemented method comprising:

grouping a first keyword into one or more first groups of related keywords;

translating, using a machine translation process and after grouping the first keyword into one or more first groups of related keywords, a first keyword from a source language into a plurality of second keywords in a target language;

after translating the first keywords into the second keywords, grouping the second keywords into one or more second groups of keywords based on the first groups and the first keyword from which each of the second keyword was translated;

identifying the one or more pre-determined word clusters for the second keywords based on the second groups of keywords;

augmenting the second keywords to include additional keywords in the target language that each have at least a threshold association with one or more of the second keywords, wherein the additional keywords are determined to have at least the threshold association with the one or more second keywords based on the additional keywords and the one or more second keywords being included in the one or more pre-determined word clusters;

determining, by a computer system after augmenting the second keywords, frequencies with which each of the second keywords occur in a corpus associated with the target language, wherein the corpus comprises one or more corpora used by a search engine, and wherein determining the frequencies includes, at least in part, conducting a search for each of the second keywords using the search engine;

selecting, by the computer system, a particular keyword from the second keywords to use in the target language based on the determined frequencies of occurrence;

associating the particular keyword in the target language with an advertisement in the target language;

receiving a search query comprising one or more search terms in the target language;

determining, by the computer system, whether to provide the advertisement with search results for the search query based, at least, on a comparison of the particular keyword and the one or more search terms of the search query; and

based on the determination of whether to provide the advertisement, providing the advertisement.

2. The computer-implemented method of claim 1, wherein the additional keywords are determined to have at least the threshold association with the one or more second keywords additionally based on the additional keywords and the one or more second keywords being identified as synonyms in the target language.

3. The computer-implemented method of claim 1, further comprising extracting one or more of the additional keywords from one or more landing pages that are in the target language and that are associated with the advertisement.

4. The computer-implemented method of claim 1, wherein the corpus further comprises search logs of search queries associated with the target language.

5. The computer-implemented method of claim 4, wherein the determined frequencies are further based on search logs for search queries received within a threshold amount of time from a current time.

6. The computer-implemented method of claim 1, wherein the corpus further comprises user-generated content associated with the target language and from one or more social networks.

7. The computer-implemented method of claim 1, further comprising:

receiving a request to translate the first keyword into a target language that is associated with a geographic region; and

selecting, from a plurality of corpora, the corpus for use in determining the frequencies of the second keywords based on the corpus being associated with the target language and the geographic region.

8. The computer-implemented method of claim 1, wherein each keyword comprises one or more words.

9. A system for translation of keywords from a source language to a target language, the system comprising:

one or more computers;

a machine translation module of the computers that is configured to translate a first keyword from a source language into a plurality of second keywords in a target language;

a keyword grouping module of the computers that is configured to i) group the first keyword into one or more first groups of related keywords, ii) after translating the first keyword into the second keywords, group the second keywords into one or more second groups of keywords based on the first groups and the first keyword from which each of the second keyword was translated, and iii) identify the one or more pre-determined word clusters for the second keywords based on the second groups of keywords;

a keyword augmentation module that is configured to augment the second keywords to include additional keywords in the target language that each have at least a threshold association with one or more of the second keywords, wherein the additional keywords are determined to have at least the threshold association with the one or more second keywords based on the additional keywords and the one or more second keywords being included in the one or more pre-determined word clusters;

a keyword frequency module of the computers that is configured to determine frequencies with which each of the second keywords occur in a corpus associated with the target language, wherein the corpus comprises one or more corpora used by a search engine, and wherein the frequencies are determined, at least in part, by conducting a search for each of the second keywords using the search engine;

a keyword selection component of the computers that is configured to select a particular keyword from the second keywords to use in the target language based on the determined frequencies of occurrence; and

a content computer system that is configured to;

associate the particular keyword in the target language with an advertisement in the target language;

receive search queries comprising one or more search terms in the target language;

determine whether to provide the advertisement with search results for the search queries based, at least, on a comparison of the particular keyword and the one or more search terms of the search queries; and

based on the determination of whether to provide the advertisement, provide the advertisement.

13. The system of claim 9, wherein the keyword augmentation module is further configured to extract one or more of the additional keywords from one or more landing pages that are in the target language and that are associated with the advertisement.

14. The system of claim 9, further comprising:

an interface of the computers to receive a request to translate the first keyword into a target language that is associated with a geographic region; and

a repository of corpora that are associated with a plurality of geographic regions and that are in a plurality of languages;

wherein the keyword frequency module is further configured to select, from the repository of corpora, the corpus for use in determining the frequencies of the second keywords based on the corpus being associated with the target language and the geographic region.

15. The system of claim 9, wherein the additional keywords are determined to have at least the threshold association with the one or more second keywords additionally based on the additional keywords and the one or more second keywords being identified as synonyms in the target language.

16. The system of claim 9, wherein the corpus further comprises search logs of search queries associated with the target language.

17. The system of claim 16, wherein the determined frequencies are further based on search logs for search queries received within a threshold amount of time from a current time.

18. The system of claim 9, wherein the corpus further comprises user-generated content associated with the target language and from one or more social networks.

19. The system of claim 9, wherein each keyword comprises one or more words.

10. An apparatus comprising a computer-readable storage medium encoding instructions thereon, that in response to execution by a computing device cause the computing device to perform operations comprising:

grouping a first keyword into one or more first groups of related keywords;

translating, using a machine translation process and after grouping the first keyword into one or more first groups of related keywords, a first keyword from a source language into a plurality of second keywords in a target language;

after translating the first keywords into the second keywords, grouping the second keywords into one or more second groups of keywords based on the first groups and the first keyword from which each of the second keyword was translated;

identifying the one or more pre-determined word clusters for the second keywords based on the second groups of keywords;

augmenting the second keywords to include additional keywords in the target language that each have at least a threshold association with one or more of the second keywords, wherein the additional keywords are determined to have at least the threshold association with the one or more second keywords based on the additional keywords and the one or more second keywords being included in the one or more pre-determined word clusters;

after augmenting the second keywords, determining frequencies with which each of the second keywords occur in a corpus associated with the target language, wherein the corpus comprises one or more corpora used by a search engine, and wherein determining the frequencies includes, at least in part, conducting a search for each of the second keywords using the search engine; and

selecting a particular keyword from the second keywords to use in the target language based on the determined frequencies of occurrence;

associating the particular keyword in the target language with an advertisement in the target language;

receiving a search query comprising one or more search terms in the target language;

determining whether to provide the advertisement with search results for the search query based, at least, on a comparison of the particular keyword and the one or more search terms of the search query; and

based on the determination of whether to provide the advertisement, providing the advertisement.

11. The apparatus of claim 10, wherein the operations further comprise extracting one or more of the additional keywords from one or more landing pages that are in the target language and that are associated with the advertisement.

12. The apparatus of claim 10, wherein the operations further comprise:

receiving a request to translate the first keyword into a target language that is associated with a geographic region; and

selecting, from a plurality of corpora, the corpus for use in determining the frequencies of the second keywords based on the corpus being associated with the target language and the geographic region.

20. The apparatus of claim 10, wherein the additional keywords are determined to have at least the threshold association with the one or more second keywords additionally based on the additional keywords and the one or more second keywords being identified as synonyms in the target language.

21. The apparatus of claim 10, wherein the corpus further comprises search logs of search queries associated with the target language.

22. The apparatus of claim 21, wherein the determined frequencies are further based on search logs for search queries received within a threshold amount of time from a current time.