CROSS-LINGUAL INFORMATION RETRIEVAL

• US 20080275691A1
• Filed: 03/28/2007
• Published: 11/06/2008
• Est. Priority Date: 01/25/2007
• Status: Active Grant

• Current Assignee: Ben Group, Inc.
• Original Assignee: Corbis Corporation
• Inventors: Uhlig, Heike, Funnell, Jarett, Summerlin, Joel, Yerigan, Wayne
• Granted Patent: US 7,933,765 B2
• Time in Patent Office: Days
• US Class Current: 704/8
• CPC Class Codes: G06F 16/3337 Translation of the query language, e.g. Chinese to English
• CPC Class Codes: G06F 40/242 Dictionaries
• CPC Class Codes: G06F 40/40 Processing or translation of natural language natural language analysis G06F40/20; semantic analysis G06F40/30

Abstract

Multi-lingual search and retrieval of digital content. Embodiments are generally directed to methods and systems for creating an English language database that associates non-English terms with English terms in multiple categories of metadata. Language experts use an interface to create equivalencies between non-English terms and English terms, Boolean expressions, synonyms, and other forms of search terms. Language dictionaries and other sources also create equivalencies. The database is used to evaluate non-English search terms submitted by a user, and to determine English search terms that can be used to perform a search for content. The multiple categories of metadata may comprise structured data, such as keywords of a structured vocabulary, and/or unstructured data, such as captions, titles, descriptions, etc. Weighting and/or prioritization can be applied to the search terms, to the process of searching the multiple categories, and/or to the search results, to rank the search results.

Claims

1. A method for identifying content, comprising:

generating an equivalency list that includes a secondary-language query term associated with at least one primary-language query term, wherein each of the at least one primary-language query term is in a pre-selected language and the secondary-language term is in a language that is different from the pre-selected primary language;

receiving the secondary-language query term in a search request;

selecting the at least one primary-language query term from the equivalency list, based on the secondary-language query term;

identifying digital content that is associated with structured metadata, if the at least one primary-language query term is included in the structured metadata; and

identifying digital content that is associated with unstructured metadata, if the at least one primary-language query term is included in the unstructured metadata and is not a unique identifier of a defined term in a controlled vocabulary.

2. The method of claim 1, wherein the equivalency list is one of a plurality of equivalency lists, each of which comprises another query term in a language different from that of any other one of the plurality of equivalency lists, and wherein the other query term is associated with the at least one primary-language query term.

3. The method of claim 1, wherein the at least one primary-language query term comprises at least one of the following;

a controlled vocabulary keyword, the secondary-language query term, a synonym associated with the secondary-language query term, and a boolean expression of terms in the pre-selected language.

4. The method of claim 1, wherein generating the equivalency list comprises:

providing an interface enabling a user to associate the at least one primary-language query term with at least one term in the language of the secondary-language term that is different from the pre-selected primary language;

receiving an indication through the interface that the secondary-language query term is a primary term; and

receiving an indication through the interface on whether a unique identifier shall be associated with a controlled vocabulary term in the at least one primary-language query term.

5. The method of claim 1, wherein generating the equivalency list comprises determining whether to remove the unique identifier from the at least one primary-language query term.

6. The method of claim 1, wherein the unique identifier indicates a limitation of a meaning of the at least one primary-language query term.

7. The method of claim 1, wherein the structured metadata comprises at least one keyword from a controlled vocabulary, wherein each of the at least one keyword is identified by a unique identifier associated with a precise concept.

8. The method of claim 1, wherein the unstructured metadata comprises at least one of the following;

a caption, a title, a paragraph, and a date.

9. The method of claim 1, further comprising weighting at least one of the following;

the at least one primary-language query term, the structured metadata, and the unstructured metadata.

10. The method of claim 1, further comprising prioritizing the identified content based on a weighting of at least one of the following;

the at least one primary-language query term, the structured metadata, and the unstructured metadata.

11. The method of claim 1, further comprising removing a noise word from the secondary-language query term prior to selecting the at least one primary-language query term from the equivalency list.

12. The method of claim 1, wherein the pre-selected language comprises English.

13. A machine readable medium storing data processing instructions that cause a machine to perform the steps of claim 1.

14. A system for identifying content, comprising:

a translations generator that performs a plurality of operations including;

generating an equivalency list that includes a secondary-language query term associated with at least one primary-language query term, wherein each of the at least one primary-language query term is in a pre-selected language and the secondary-language term is in a language that is different from the pre-selected primary language;

receiving the secondary-language query term in a search request; and

selecting the at least one primary-language query term from the equivalency list, based on the secondary-language query term; and

a search engine that performs a plurality of operations, including;

identifying digital content that is associated with structured metadata, if the at least one primary-language query term is included in the structured metadata; and

identifying digital content that is associated with unstructured metadata, if the at least one primary-language query term is included in the unstructured metadata and is not a unique identifier of a defined term in a controlled vocabulary.

15. The system of claim 14, further comprising a client that performs a plurality of operations, including:

providing an interface enabling a user to associate the at least one primary-language query term with at least one term in the language of the secondary-language term that is different from the pre-selected primary language;

receiving an indication through the interface that the secondary-language query term is a primary term; and

receiving an indication through the interface on whether a unique identifier shall be associated with a controlled vocabulary term in the at least one primary-language query term.

16. The system of claim 14, wherein the search engine further performs the operation of prioritizing the identified content based on a weighting of at least one of the following;

the at least one primary-language query term, the structured metadata, and the unstructured metadata.

17. A method for associating terms for identifying content;

comprising;

associating a secondary-language term with a controlled vocabulary keyword in a primary-language, if the secondary-language term has a unique meaning depending on a context;

indicating that the secondary-language term exists in the primary language, if the secondary-language term is identical in the primary language;

associating the secondary-language term with a synonym in the primary language, if the secondary-language term is synonymous with the synonym; and

associating the secondary-language term with a Boolean expression, if a meaning of the secondary-language term can be expressed by a combination of primary-language terms.

18. The method of claim 17, wherein the controlled vocabulary keyword includes a unique identifier indicating that the controlled vocabulary keyword has a meaning depending on the context.

19. The method of claim 17, further comprising at least one of the following:

identifying digital content associated with the secondary-language term based on structured metadata that is associated with the controlled vocabulary keyword; and

identifying digital content associated with the secondary-language term based on unstructured metadata that is associated with at least one of the following;

the secondary-language term that is identical in the primary language, the synonym in the primary language; and

the Boolean expression.

20. A method for generating a list for identifying content, comprising:

receiving a subset of equivalencies comprising a plurality of secondary-language terms that are associated with a primary-language term;

parsing the subset into a list of equivalencies, wherein each equivalency comprises an association of at least one of the plurality of secondary-language terms with the primary-language term;

associating a unique identifier with the primary-language term in at least one equivalency of the list, if at least one of the plurality of secondary-language terms has a limited meaning that is associated with the unique identifier;

adding at least one of the secondary-language terms to at least one equivalency in the list, if the primary-language term is identical to the at least one secondary-language term, wherein the at least one secondary-language term is one of the plurality of secondary-language terms;

adding a primary-language lead-in term to at least one equivalency in the list, if at least one of the plurality of secondary-language terms is synonymous with the primary-language lead-in term; and

adding a Boolean expression to at least one equivalency in the list, if the least one of the plurality of secondary-language terms is associated with a combination of terms in the primary language.

21. The method of claim 20, wherein one of the plurality of secondary-language terms is designated as a primary translation of the primary-language term into the secondary language.

22. The method of claim 20, wherein the subset of equivalencies is received from an interface that enables a user to designate associations between the plurality of secondary-language terms and the primary-language term.

23. The method of claim 20, further comprising providing the list of equivalencies to a search engine that identifies digital content based on the list and at least one of the following;

structured metadata and unstructured metadata.

24. A system for generating a list for identifying content, comprising:

a parser that performs a plurality of operations, including;

receiving a subset of equivalencies comprising a plurality of secondary-language terms that are associated with a primary-language term; and

parsing the subset into a list of equivalencies, wherein each equivalency comprises an association of at least one of the plurality of secondary-language terms with the primary-language term; and

a list generator that performs a plurality of operations, including;

associating a unique identifier with the primary-language term in at least one equivalency of the list, if at least one of the plurality of secondary-language terms has a limited meaning that is associated with the unique identifier;

adding a nonprimary-language term to at least one equivalency in the list, if the primary-language term is identical to the nonprimary-language term, wherein the nonprimary-language term is one of the plurality of secondary-language terms;

adding a primary-language lead-in term to at least one equivalency in the list, if at least one of the plurality of secondary-language terms is synonymous with the primary-language lead-in term; and

adding a Boolean expression to at least one equivalency in the list, if the least one of the plurality of secondary-language terms is associated with a combination of terms in the primary language.

25. The system of claim 24 wherein the list generator further performs the operation of providing the list of equivalencies to a search engine that identifies digital content based on the list and at least one of the following:

structured metadata and unstructured metadata.

26. A method for determining a query to identify content, comprising:

receiving a first equivalency between;

a primary-language query term in a primary language; and

a user-specified secondary-language query term in a secondary language;

receiving a second equivalency between the primary-language query term and an alternate secondary-language query term in the secondary language;

determining whether to apply a unique identifier to either of the user-specified secondary-language query term or the alternate secondary-language query term, wherein the unique identifier refines the meaning of a query term and indicates a structured query term;

receiving a search query in the secondary language; and

determining the primary-language query term based at least in part on the search query, the user-specified secondary-language query term, and the alternate secondary-language query term.

27. The method of claim 26, wherein the primary language is English and the secondary language is one of a plurality of languages other than English.

28. The method of claim 26, wherein the alternate secondary language query term comprises at least one of the following;

a dictionary entry, a user-defined Boolean expression, and a synonym.

29. The method of claim 26, further comprising providing the primary-language query term to a search engine for searching at least one of the following:

structured metadata, which comprises a controlled vocabulary of keywords associated with content; and

unstructured metadata, which comprises free-text associated with content, wherein the free-text comprises categories of text that need not conform to a controlled vocabulary.

30. The method of claim 26, further comprising removing a noise word from at least one of the following;

the user-specified secondary-language query term and the alternate secondary-language query term.