Query language determination using query terms and interface language

• US 8,762,358 B2
• Filed: 04/19/2006
• Issued: 06/24/2014
• Est. Priority Date: 04/19/2006
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Datta, Ruchira S., Lopiano, Fabio
• Primary Examiner(s): Smith, Brannon W
• Application Number:
  US11/408,245
• Publication Number:
  US 20070288450A1
• Time in Patent Office: 2,988 Days
• Field of Search: 707/4, 707/999.004, 707/748, 707/706, 707/728
• US Class Current: 707/706
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 40/126 Character encoding
• CPC Class Codes: G06F 40/129 Handling non-Latin characters, e.g. kana-to-kanji conversion
• CPC Class Codes: G06F 40/247 Thesauruses; Synonyms

Abstract

Methods, systems, and apparatus, including computer program products, operable to perform operations including receiving from a user through a user interface a search query comprising one or more query terms, the user interface having an interface language, the interface language being a natural language; and determining a query language for the query from the query terms and the interface language, the query language being a natural language.

Claims

1. A computer implemented method comprising:

determining, using one or more data processing apparatus, an interface language associated with a user interface through which one or more query terms of a search query are received;

identifying respective languages that are associated with search result selections for a plurality of previously submitted search queries, each of the previously submitted search queries matching the search query;

determining, for each identified language and based on the respective languages that are associated with the search result selections, a respective likelihood that the search query is in the language; and

selecting a likelihood for a language that matches the interface language as a likelihood that one or more of the query terms are in the interface language;

determining, for each of the query terms, a vector corresponding to a set of likelihoods that the respective query term is in a plurality of languages;

multiplying together the vectors corresponding to the sets of likelihoods that the query terms are in the plurality of languages to generate a resultant vector;

multiplying the resultant vector by the likelihood that one or more of the query terms are in the determined interface language to generate a query probability vector; and

selecting a query language of the query based at least on the query probability vector.

2. The method of claim 1, further comprising:

identifying, based on a language-specific mapping, a respective common form of the one or more query terms for one or more of the one or more query terms;

identifying, for the one or more query terms, a respective synonym based on the respective common forms of the one or more query terms; and

revising the search query based on the one or more identified synonyms.

3. The method of claim 2, wherein the language-specific mapping comprises a collection of one or more character conversion maps, the character conversion maps specifying one or more input characters and one or more output characters to which the one or more input characters are mapped.

4. The method of claim 2, wherein identifying, for the one or more query terms, a respective synonym based on the respective common forms of the one or more query terms comprises:

identifying, for a term in the one or more query terms, one or more potential synonyms;

determining, for each synonym in the one or more potential synonyms, a relative synonym probability for each language in the plurality of languages; and

selecting, based on the synonym probabilities, a synonym from the one or more potential synonyms.

5. The method of claim 4, wherein determining, for each synonym in the one or more potential synonyms, a relative synonym probability for each language in the plurality of languages comprises:

identifying, for a synonym in the one or more potential synonyms, a relative frequency of the synonym among documents in each language in the plurality of languages;

computing, for each language in the plurality of languages, a language estimate, wherein the language estimate is based on the relative frequency of the synonym for the language and the likelihood that the one or more query terms are in the language; and

aggregating the language estimates for each language in the plurality of languages to determine a synonym probability for the synonym.

6. A computer program product, encoded on a non-transitory machine-readable storage device, operable to cause data processing apparatus to perform operations comprising:

determining, using one or more data processing apparatus, an interface language associated with a user interface through which one or more query terms of a search query are received;

identifying respective languages that are associated with search result selections for a plurality of previously submitted search queries, each of the previously submitted search queries matching the search query;

determining, for each identified language and based on the respective languages that are associated with the search result selections, a respective likelihood that the search query is in the language; and

selecting a likelihood for a language that matches the interface language as a likelihood that one or more of the query terms are in the interface language;

determining, for each of the query terms, a vector corresponding to a set of likelihoods that the respective query term is in a plurality of languages;

multiplying together the vectors corresponding to the sets of likelihoods that the query terms are in the plurality of languages to generate a resultant vector;

multiplying the resultant vector by the likelihood that one or more of the query terms are in the determined interface language to generate a query probability vector; and

selecting a query language of the query based at least on the query probability vector.

7. The program product of claim 6, further comprising:

identifying, based on a language-specific mapping, a respective common form of the one or more query terms for one or more of the one or more query terms;

identifying, for the one or more query terms, a respective synonym based on the respective common forms of the one or more query terms; and

revising the search query based on the one or more identified synonyms.

8. The program product of claim 7, wherein the language-specific mapping comprises a collection of one or more character conversion maps, the character conversion maps specifying one or more input characters and one or more output characters to which the one or more input characters are mapped.

9. The program product of claim 7, wherein identifying, for the one or more query terms, a respective synonym based on the respective common forms of the one or more query terms comprises:

identifying, for a term in the one or more query terms, one or more potential synonyms;

determining, for each synonym in the one or more potential synonyms, a relative synonym probability for each language in the plurality of languages; and

selecting, based on the synonym probabilities, a synonym from the one or more potential synonyms.

10. The program product of claim 9, wherein determining, for each synonym in the one or more potential synonyms, a relative synonym probability for each language in the plurality of languages comprises:

identifying, for a synonym in the one or more potential synonyms, a relative frequency of the synonym among documents in each language in the plurality of languages;

computing, for each language in the plurality of languages, a language estimate, wherein the language estimate is based on the relative frequency of the synonym for the language and the likelihood that the one or more query terms are in the language; and

aggregating the language estimates for each language in the plurality of languages to determine a synonym probability for the synonym.

11. A system comprising:

one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising;

determining an interface language associated with a user interface through which one or more query terms of a search query are received;

identifying respective languages that are associated with search result selections for a plurality of previously submitted search queries, each of the previously submitted search queries matching the search query;

determining, for each identified language and based on the respective languages that are associated with the search result selections, a respective likelihood that the search query is in the language; and

selecting a likelihood for a language that matches the interface language as a likelihood that one or more of the query terms are in the interface language;

determining, for each of the query terms, a vector corresponding to a set of likelihoods that the respective query term is in a plurality of languages;

multiplying together the vectors corresponding to the sets of likelihoods that the query terms are in the plurality of languages to generate a resultant vector;

multiplying the resultant vector by the likelihood that one or more of the query terms are in the determined interface language to generate a query probability vector; and

selecting a query language of the query based at least on the query probability vector.

12. The system of claim 11, further comprising:

identifying, based on a language-specific mapping, a respective common form of the one or more query terms for one or more of the one or more query terms;

identifying, for the one or more query terms, a respective synonym based on the respective common forms of the one or more query terms; and

revising the search query based on the one or more identified synonyms.

13. The system of claim 12, wherein the language-specific mapping comprises a collection of one or more character conversion maps, the character conversion maps specifying one or more input characters and one or more output characters to which the one or more input characters are mapped.

14. The system of claim 12, wherein identifying, for the one or more query terms, a respective synonym based on the respective common forms of the one or more query terms comprises:

identifying, for a term in the one or more query terms, one or more potential synonyms;

determining, for each synonym in the one or more potential synonyms, a relative synonym probability for each language in the plurality of languages; and

selecting, based on the synonym probabilities, a synonym from the one or more potential synonyms.

15. The system of claim 14, wherein determining, for each synonym in the one or more potential synonyms, a relative synonym probability for each language in the plurality of languages comprises:

identifying, for a synonym in the one or more potential synonyms, a relative frequency of the synonym among documents in each language in the plurality of languages;

computing, for each language in the plurality of languages, a language estimate, wherein the language estimate is based on the relative frequency of the synonym for the language and the likelihood that the one or more query terms are in the language; and

aggregating the language estimates for each language in the plurality of languages to determine a synonym probability for the synonym.