Determining a meaning of a knowledge item using document based information

• US 8,433,671 B2
• Filed: 04/11/2011
• Issued: 04/30/2013
• Est. Priority Date: 09/22/1999
• Status: Expired due to Term

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Elbaz, Gilad Israel, Weissman, Adam J.
• Primary Examiner(s): Chaki, Kakali
• Assistant Examiner(s): COUGHLAN, PETER D
• Application Number:
  US13/084,377
• Publication Number:
  US 20110191175A1
• Time in Patent Office: 750 Days
• Field of Search: 706/55
• US Class Current: 706/55
• CPC Class Codes: G06F 16/313 Selection or weighting of terms for indexing
• CPC Class Codes: G06F 16/3338 Query expansion
• CPC Class Codes: G06F 16/36 Creation of semantic tools, e.g. ontology or thesauri
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06Q 30/0251 Targeted advertisements

Abstract

Systems and methods that determine a meaning of a knowledge item using related information are described. In one aspect, a knowledge item is received, related information associated with the knowledge item is received, at least one related meaning based on the related information is determined, and a knowledge item meaning for the knowledge item based at least in part on the related meaning is determined. Several algorithms and types of related information useful in carrying out such systems and methods are described.

Claims

1. A method performed by a computer, the method comprising:

receiving a knowledge item, the knowledge item including knowledge item text;

determining one or more knowledge item concepts associated with the knowledge item, each knowledge item concept being represented by one or more weighted items;

receiving one or more related information items that are each associated with the knowledge item, wherein the one or more related information items include related information text;

determining related information text meanings, wherein each of the related information text meanings is a meaning of at least a portion of the related information text and is represented by one or more weighted terms;

calculating, for each knowledge item concept and each related information text meaning, a strength of a relationship between the knowledge item concept and the related information text meaning;

selecting a particular knowledge item concept based on the strengths of the relationships; and

designating the selected particular knowledge item concept as a meaning of the knowledge item.

2. The method of claim 1, wherein none of the one or more related information items is a part of the knowledge item.

3. The method of claim 1, wherein the knowledge item includes at least one word.

4. The method of claim 1, wherein the one or more related information items include one or more advertisements, each advertisement being related to the knowledge item through acts of bidding.

5. The method of claim 1, wherein each of the knowledge item concepts includes at least one keyword concept, the keyword concept being represented by one or more weighted keyword terms.

6. The method of claim 5, wherein the strength of relationship between the knowledge item concept and the related information text meaning indicates a probability of co-occurrence between the keyword concept and a meaning concept.

7. The method of claim 6, wherein calculating the strength of relationship between the knowledge item concept and the related information text meaning includes adjusting the probability of co-occurrence based on weights of keyword terms representing the keyword concept and weights of terms representing the meaning concept.

8. A method for determining a meaning of a keyword, the method comprising:

receiving, in a system comprising one or more computers, a keyword;

receiving, in the system, text of advertisements associated with advertisers who have currently bid on the keyword;

determining, in the system, a plurality of related meanings from the text of the advertisements associated with the advertisers who have currently bid on the keyword;

obtaining, in the system, a plurality of previously stored candidate meanings of the keyword;

calculating, by the system, a strength of relationship between each of the plurality of candidate meanings and each of the plurality of related meanings;

selecting, by the system, a candidate meaning from the plurality of candidate meanings based on the strengths of relationship; and

designating, by the system, the selected candidate meaning as a meaning of the keyword.

9. The method of claim 8, wherein:

each of the related meanings and each of the candidate meanings including the selected candidate meaning is represented in the system as a respective vector of weighted concepts or as a respective cluster of related words.

10. The method of claim 8, wherein:

the text of the advertisements also includes additional related data.

11. The method of claim 8, wherein:

the additional related data includes data related to a success rate of the advertisements.

12. The method of claim 8, wherein:

the additional related data includes data related to other keywords purchased by advertisers of the advertisements.

13. The method of claim 8, further comprising:

determining that the keyword comprises a first term and a second term, and that a set of first concepts is associated with the first term and a set of second concepts is associated with the second term;

determining that the first concepts and the second concepts include a connected concept; and

increasing a weight of the connected concept in a vector of weighted concepts representing a candidate meaning for the keyword.

14. 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;

receiving, in a system comprising one or more computers, a keyword;

receiving, in the system, text of advertisements associated with advertisers who have currently bid on the keyword;

determining, in the system, a plurality of related meanings from the text of the advertisements associated with the advertisers who have currently bid on the keyword;

obtaining, in the system, a plurality of previously stored candidate meanings of the keyword;

calculating, by the system, a strength of relationship between each of the plurality of candidate meanings and each of the plurality of related meanings;

selecting, by the system, a candidate meaning from the plurality of candidate meanings based on the strengths of relationship; and

designating, by the system, the selected candidate meaning as a meaning of the keyword.

15. The system of claim 14, wherein:

each of the related meanings and each of the candidate meanings including the selected candidate meaning is represented in the system as a respective vector of weighted concepts or as a respective cluster of related words.

16. The system of claim 14, wherein:

the text of the advertisements also includes additional related data.

17. The system of claim 14, wherein:

the additional related data includes data related to a success rate of the advertisements.

18. The system of claim 14, wherein:

the additional related data includes data related to other keywords purchased by advertisers of the advertisements.

19. The system of claim 14, wherein the operations further comprise:

determining that the keyword comprises a first term and a second term, and that a set of first concepts is associated with the first term and a set of second concepts is associated with the second term;

determining that the first concepts and the second concepts include a connected concept; and

increasing a weight of the connected concept in a vector of weighted concepts representing a candidate meaning for the keyword.

20. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:

receiving, in a system comprising one or more computers, a keyword;

receiving, in the system, text of advertisements associated with advertisers who have currently bid on the keyword;

determining, in the system, a plurality of related meanings from the text of the advertisements associated with the advertisers who have currently bid on the keyword;

obtaining, in the system, a plurality of previously stored candidate meanings of the keyword;

calculating, by the system, a strength of relationship between each of the plurality of candidate meanings and each of the plurality of related meanings;

selecting, by the system, a candidate meaning from the plurality of candidate meanings based on the strengths of relationship; and

designating, by the system, the selected candidate meaning as a meaning of the keyword.

21. The medium of claim 20, wherein:

each of the related meanings and each of the candidate meanings including the selected candidate meaning is represented in the system as a respective vector of weighted concepts or as a respective cluster of related words.

22. The medium of claim 20, wherein:

the text of the advertisements also includes additional related data.

23. The medium of claim 20, wherein:

the additional related data includes data related to a success rate of the advertisements.

24. The medium of claim 20, wherein:

the additional related data includes data related to other keywords purchased by advertisers of the advertisements.

25. The medium of claim 20, wherein the operations further comprise:

determining that the keyword comprises a first term and a second term, and that a set of first concepts is associated with the first term and a set of second concepts is associated with the second term;

determining that the first concepts and the second concepts include a connected concept; and

increasing a weight of the connected concept in a vector of weighted concepts representing a candidate meaning for the keyword.