Document processing

• US 7,707,206 B2
• Filed: 09/21/2006
• Issued: 04/27/2010
• Est. Priority Date: 09/21/2005
• Status: Expired due to Fees

• Current Assignee: Praxeon Incorporated
• Original Assignee: Praxeon Incorporated
• Inventors: Underwood, Dennis, Gilpin, Kevin, Encina, Nicolas
• Primary Examiner(s): Rones; Charles
• Assistant Examiner(s): Khoshnoodi; Fariborz
• Application Number:
  US11/525,245
• Publication Number:
  US 20080140616A1
• Time in Patent Office: 1,314 Days
• Field of Search: 703/3, 700/1, 713/168, 358/1.13, 435/7.23, 707/3, 707/5, 707/4
• US Class Current: 707/716
• CPC Class Codes: G06F 16/31 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

A search system and method are disclosed. In one general aspect the method features fingerprinting digitally stored documents and queries based on one or more topic models. Similarities between document and query fingerprints can be detected to search for documents. Topic summaries can also be derived from the documents and queries based on the topic models.

Claims

1. A computerized search system comprising a processor and a storage device including instructions that are configured to run on the processor for accessing digitally stored data, comprising:

a document interface configured to access a plurality of digitally stored documents that each includes content that relates to one or more subject matter topics,

a topic model interface configured to access one or more digitally stored topic models that each classify information about one of the subject matter topics that can occur in the content, wherein the topic model interface is configured to access topic models that include hierarchical concept maps that map each of one or more ancestor concepts to a plurality of descendent concepts,

document fingerprinting logic embodied in the computerized search system and responsive to the digitally stored topic models through the topic model interface and to the digitally stored documents through the document interface, and configured to create document fingerprints that each include a set of identifiers that each identify one of the topics from the digitally stored topic models in the content of the digitally stored documents,

a query interface configured to receive user-specified queries,

query fingerprinting logic embodied in the computerized search system and responsive to the digitally stored topic models through the topic model interface and to queries through the query interface, and configured to create query fingerprints that identify topics from the stored topic models in the queries, and

search logic embodied in the computerized search system and configured to identify one or more of the digitally stored documents that are relevant to the queries, based on the query fingerprints and the document fingerprints.

2. The system of claim 1 wherein the search logic includes an indexer configured to index the document fingerprints, and a search engine configured to search the index based on the query fingerprints.

3. The system of claim 2 further including analysis and augmentation logic configured to modify the document and/or query fingerprints before processing by the indexer or search engine.

4. The system of claim 3 wherein the analysis and augmentation logic is responsive to curation input.

5. The system of claim 3 wherein the analysis and augmentation logic includes context-sensitive meaning scope adjustment logic configured to expand or restrict topics identified by a term in a fingerprint to a set of the meanings associated with that term.

6. The system of claim 3 wherein the analysis and augmentation logic includes species augmentation logic configured to add specific examples of a generic topic.

7. The system of claim 6 wherein the species augmentation logic operates based on chemical similarity or biological sequence similarity.

8. The system of claim 6 wherein the species augmentation logic operates based on predefined hierarchical relationships between topics.

9. The system of claim 2 further including a ranker responsive to the stored topic models and configured to rank topics from a set of fingerprints or search engine results.

10. The system of claim 2 further including a result presentation interface responsive to the search engine and configured to present results to the user in a manner that is based on the document fingerprints.

11. The system of claim 1 further including an alert interface configured to generate alerts based on the fingerprints.

12. The system of claim 1 wherein the topic models include topical models and location models.

13. The system of claim 1 wherein the topic models include topical models and taxonomic models.

14. The system of claim 1 wherein the topic models include topical models and biographical models.

15. The system of claim 1 wherein the topic models include topical models and historical models.

16. The system of claim 1 wherein the topic models include topical models and disease diagnosis models.

17. The system of claim 1 wherein the topic models include topical models and disease treatment models.

18. The system of claim 1 wherein the topic models include topical models and chemical models.

19. The system of claim 1 wherein the topic models include topical models and genetic models.

20. The system of claim 1 wherein the topic models include medical treatment and diagnosis models.

21. The system of claim 1 wherein the topic models include topical models and industrial classification models.

22. The system of claim 1 wherein the query fingerprinting logic and the document fingerprinting logic operate in the same way.

23. The system of claim 1 further including a topical model update interface configured to update the topical models.

24. The system of claim 23 wherein the topical model update interface is responsive to curation input.

25. The system of claim 23 further including a topical model version management system configured to selectively update a subset of document fingerprints in response to changes to the stored topical models.

26. The system of claim 1 wherein the query fingerprinting logic is configured to generate fingerprints based on free-form naturally posed queries.

27. The system of claim 1 wherein the query interface is part of a syndication filtering tool.

28. The system of claim 27 further including a feedback mechanism responsive to user feedback to update a syndication query fingerprint.

29. The system of claim 27 wherein the query interface is part of a syndication filtering tool that includes a plug-in interface to a document access tool.

30. The system of claim 29 further including a feedback mechanism responsive to user feedback to update a syndication query fingerprint for the syndication tool through the plug-in interface.

31. The system of claim 1 further including classification logic configured to assist in locating documents in predetermined classes.

32. The system of claim 1 wherein at least one of the topic models includes synonym tables that can narrow or expand the range of synonyms associated with a term in the context of that model.

33. The system of claim 1 further including topic-based post-search processing logic responsive to the topic models and the search logic, and configured to selectively access, based on the topic models, information from result sets produced by the search logic.

34. A computer implemented method, employing a processor and a storage device including instructions that are configured to run on the processer, comprising:

accessing digitally stored documents that each include that relates to one or more subject matter topics,

accessing one or more topic models that each classify information about one of the subject matter topics that can occur in the content, wherein the step of accessing accesses topic models that include hierarchical concept maps that map each of one or more ancestor concepts to a plurality of descendent concepts,

fingerprinting the documents accessed in the step of accessing documents based on the topic models accessed in the step of accessing topic models,

accessing one or more user-specified queries,

fingerprinting the queries accessed in the step of accessing queries based on the topic models accessed in the step of accessing topic models, and

detecting similarities between fingerprints produced by the step of fingerprinting documents and the step of fingerprinting queries.

35. A computerized search system, comprising a processor and a storage device including instructions that are configured to run on the processor for accessing digitally stored data, comprising:

means for accessing digitally stored documents that each content that relates to one or more subject matter topics,

means for accessing one or more topic models that each classify information about one of the subject matter topics that can occur in the content,

means for fingerprinting documents accessed by the means for accessing documents based on the topic models accessed by the means for accessing topic models,

means for accessing one or more user-specified queries, wherein the means for accessing accesses topic models that include hierarchical concept maps that map each of one or more ancestor concepts to a plurality of descendent concepts,

means for fingerprinting the queries accessed by the means for accessing queries based on the topic models accessed by the means for accessing topic models, and

means for detecting similarities between fingerprints produced by the means for fingerprinting documents and the means for fingerprinting queries.