Augmenting queries with synonyms from synonyms map

• US 8,606,826 B2
• Filed: 01/12/2012
• Issued: 12/10/2013
• Est. Priority Date: 04/19/2006
• Status: Expired due to Fees

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Datta, Ruchira S.
• Primary Examiner(s): Chbouki, Tarek
• Application Number:
  US13/349,433
• Publication Number:
  US 20120109978A1
• Time in Patent Office: 698 Days
• Field of Search: None
• US Class Current: 707/803
• CPC Class Codes: G06F 16/3338 Query expansion
• CPC Class Codes: G06F 40/247 Thesauruses; Synonyms

Abstract

Methods, systems, and apparatus, including computer program products, operable to perform operations including receiving through a user interface with an interface language a search query having query terms; using the interface language to select one or more mappings and using the selected mappings to simplify each query term; and applying each simplified query term to a synonyms map to identify possible synonyms with which to augment the search query. In alternative embodiments, the operations include generating a synonyms map from a corpus of documents; where the synonyms map maps each of multiple keys to one or more corresponding variants, where each variant is associated with one or more of document languages. In alternative embodiments, the operations include generating a synonyms map from documents by applying document language-dependent mappings to words in the documents to generate keys for the map.

Claims

1. A computer implemented method, comprising:

generating, in a system comprising one or more computers, a synonyms map from a corpus of documents, wherein;

each document is associated with a document language, the document language representing a natural language of the document,

the synonyms map maps each of a plurality of keys to one or more corresponding variants,

each key is a common form word whose characters do not include diacritical marks,

each of the one or more corresponding variants is a variant of the common form word each variant (i) being found in the corpus of documents, (ii) being determined based on one or more character conversion maps that each specify one or more output characters to which one or more corresponding input characters are mapped, and (iii) including one or more characters that include diacritical marks; and

the synonyms map associates each variant of a key with;

(i) two or more document languages, and (ii) a respective score for each of the two or more document languages, the score indicating a relative frequency of the variant in documents associated with the document language among all variants of the key for the document language.

2. The method of claim 1, wherein:

the attribution of a document language to each document is determined automatically.

3. The method of claim 1 wherein the corresponding variants are stored in a trie.

4. The method of claim 1, wherein the synonyms map associates each variant of a key with two or more of the document languages.

5. A non-transitory computer-readable storage device storing instructions that, when executed by a computer system, cause the computer system to perform operations comprising:

generating, in a system comprising one or more computers, a synonyms map from a corpus of documents, wherein;

each document is associated with a document language, the document language representing a natural language of the document,

each document language is a natural language,

the synonyms map maps each of a plurality of keys to one or more corresponding variants,

each key is a common form word whose characters do not include diacritical marks,

each of the one or more corresponding variants is a variant of the common form word each variant (i) being found in the corpus of documents, (ii) being determined based on one or more character conversion maps that each specify one or more output characters to which one or more corresponding input characters are mapped, and (iii) including one or more characters that include diacritical marks; and

the synonyms map associates each variant of a key with;

(i) two or more document languages, and (ii) a respective score for each of the two or more document languages, the score indicating a relative frequency of the variant in documents associated with the document language among all variants of the key for the document language.

6. The non-transitory computer-readable storage device of claim 5, wherein:

the attribution of a document language to each document is determined automatically.

7. The non-transitory computer-readable storage device of claim 5 wherein the corresponding variants are stored in a trie.

8. The non-transitory computer-readable storage device of claim 5, wherein the synonyms map associates each variant of a key with two or more of the document languages.

9. A system comprising:

one or more processors; and

a computer-readable storage device storing instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising;

generating, in a system comprising one or more computers, a synonyms map from a corpus of documents, wherein;

each document is associated with a document language, the document language representing a natural language of the document, each document language is a natural language,

the synonyms map maps each of a plurality of keys to one or more corresponding variants,

each key is a common form word whose characters do not include diacritical marks,

each of the one or more corresponding variants is a variant of the common form word each variant (i) being found in the corpus of documents, (ii) being determined based on one or more character conversion maps that each specify one or more output characters to which one or more corresponding input characters are mapped, and (iii) including one or more characters that include diacritical marks; and

the synonyms map associates each variant of a key with;

(i) two or more document languages, and (ii) a respective score for each of the two or more document languages, the score indicating a relative frequency of the variant in documents associated with the document language among all variants of the key for the document language.

10. The system of claim 9, wherein:

the attribution of a document language to each document is determined automatically.

11. The system of claim 9 wherein the corresponding variants are stored in a trie.

12. The system of claim 9, wherein the synonyms map associates each variant of a key with two or more of the document languages.