Generating descriptions of matching resources based on the kind, quality, and relevance of available sources of information about the matching resources

• US 8,745,044 B2
• Filed: 07/11/2008
• Issued: 06/03/2014
• Est. Priority Date: 09/17/2002
• Status: Expired due to Fees

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Carson, Chad, Nibhanupudi, Mohan V., Meyers, Robert, Pavlovski, Dmitri, Cook, Douglas
• Primary Examiner(s): Beausoliel, Jr., Robert
• Assistant Examiner(s): BROMELL, ALEXANDRIA Y
• Application Number:
  US12/172,165
• Publication Number:
  US 20080270394A1
• Time in Patent Office: 2,153 Days
• Field of Search: 707/708, 707/711, 707/694, 707/728, 707/769
• US Class Current: 707/728
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

Techniques are provided for generating descriptions of matching resources in a manner that takes into account the kind, quality, and relevance of the available sources of information about the matching resources. For example, after the search engine identifies matching resources based on the query terms, the search engine determines the kinds of available sources of information about each matching resource. For each matching resource, based on the kinds of available sources of information about the matching resource, one of a plurality of processes is selected to generate a description for the matching resource. Using the content-sensitive description generation techniques described herein, a single result set may include abstracts that were generated using several different processes, where the difference in process corresponds to a difference in the kind, quality, and relevance of the available sources of information about each matching resource.

Claims

1. A method of generating a composite abstract for a document, the method comprising:

receiving a query that includes one or more search query terms;

generating a first version of a contextual abstract of the document, at least in part, by selecting, based on the one or more search query terms, first portions of the document for inclusion within the contextual abstract of the document;

generating a static abstract of the document, at least in part, by selecting, without regard to the one or more search query terms, second portions of the document for inclusion within the static abstract of the document;

generating a second version of the contextual abstract by removing one or more text strings from the first version of the contextual abstract;

wherein removing the one or more text strings includes removing a particular text string from the first version of the contextual abstract based, at least in part, on similarities between the contextual abstract and the static abstract;

combining text strings from the second version of the contextual abstract with text strings from the static abstract to form the composite abstract for the document;

responding to the query with search results that include the composite abstract.

2. The method of claim 1, wherein the particular text string is a first particular text string, and further comprising:

prior to combining text strings from the second version of the contextual abstract with text strings from the static abstract, removing the one or more text strings from the static abstract;

wherein removing the one or more text string s from the static abstract includes removing a second particular text string from the static abstract based, at least in part, on similarities between the static abstract and the second version of the contextual abstract.

3. The method of claim 1, wherein similarity is based, at least in part, on a proportion of (a) words that are contained in both the particular text string and the static abstract of the document, relative to (b) words that are contained in the particular text string but not in the static abstract of the document.

4. The method of claim 1, wherein each text string in the contextual abstract satisfies specified criteria.

5. The method of claim 1, further comprising removing a part of the contextual abstract, wherein the part is contained in the static abstract.

6. The method of claim 2, further comprising removing the second particular text string from the static abstract if words from the second particular text string that occur in the static abstract comprise less than a specified percentage of words that are contained in the second particular text string.

7. An apparatus for generating a composite abstract for a document, the apparatus comprising:

a processor;

a module capable of receiving a query that includes one or more search query terms;

a module capable of generating a first version of a contextual abstract of the document, at least in part, by selecting, based on the one or more search query terms, first portions of the document for inclusion within the contextual abstract of the document;

a module capable of generating a static abstract of the document, at least in part, by selecting, without regard to the search query terms, second portions of the document for inclusion within the static abstract of the document;

a module capable of generating a second version of the contextual abstract by removing one or more text string from the first version of the contextual abstract;

wherein removing one or more text strings includes removing a particular text string from the first version of the contextual abstract based, at least in part, on similarities between the contextual abstract and the static abstract;

a module capable of combining text strings from the second version of the contextual abstract with text string from the static abstract to form the composite abstract for the document;

a module capable of responding to the query with search results that include the composite abstract.

8. The apparatus of claim 7, wherein the particular text string is a first particular text string, and further comprising:

prior to combining text strings from the second version of the contextual abstract with text strings from the static abstract, removing the one or more text strings from the static abstract;

wherein removing the one or more text strings from the static abstract includes removing a second particular text string from the static abstract based, at least in part, on similarities between the static abstract and the second version of the contextual abstract.

9. The apparatus of claim 8, further comprising a module capable of removing the second particular text string from the static abstract if words from the second particular text string that occur in the static abstract comprise less than a specified percentage of words that are contained in the second particular text string.

10. The apparatus of claim 7, wherein similarity is based, at least in part, on a proportion of (a) words that are contained in both the particular text string and the static abstract of the document, relative to (b) words that are contained in the particular text string but not in the static abstract of the document.

11. The apparatus of claim 7, wherein each text string in the contextual abstract satisfies specified criteria.

12. The apparatus of claim 7, further comprising a module capable of removing a part of the contextual abstract, wherein the part is contained in the static abstract.

13. One or more non-transitory computer-readable media storing instructions for performing a method of generating a composite abstract for a document, wherein method comprises:

receiving a query that includes one or more search query terms;

generating a first version of a contextual abstract of the document, at least in part, by selecting, based on the one or more search query terms, first portions of the document for inclusion within the contextual abstract of the document;

generating a static abstract of the document, at least in part, by selecting, without regard to the one or more search query terms, second portions of the document for inclusion within the static abstract of the document;

generating a second version of the contextual abstract by removing one or more text strings from the first version of the contextual abstract;

wherein removing the one or more text strings includes removing a particular text string from the first version of the contextual abstract based, at least in part, on similarities between the contextual abstract and the static abstract;

combining text strings from the second version of the contextual abstract with text strings from the static abstract to form the composite abstract for the document;

responding to the query with search results that include the composite abstract.

14. The one or more non-transitory computer-readable media of claim 13, wherein the particular text string is a first particular text string, and further comprising:

prior to combining text strings from the second version of the contextual abstract with text strings from the static abstract, removing the one or more text strings from the static abstract;

wherein removing the one or more text strings from the static abstract includes removing a second particular text string from the static abstract based, at least in part, on similarities between the static abstract and the second version of the contextual abstract.

15. The one or more non-transitory computer-readable media of claim 14, wherein the method further comprises removing the second particular text string from the static abstract if words from the second particular text string that occur in the static abstract comprise less than a specified percentage of words that are contained in the second particular text string.

16. The one or more non-transitory computer-readable media of claim 13, wherein similarity is based, at least in part, on a proportion of (a) words that are contained in both the particular text string and the static abstract of the document, relative to (b) words that are contained in the particular text string but not in the static abstract of the document.

17. The one or more non-transitory computer-readable media of claim 13, wherein each text string in the contextual abstract satisfies specified criteria.

18. The one or more non-transitory computer-readable media of claim 13, wherein the method further comprises removing a part of the contextual abstract, wherein the part is contained in the static abstract.