Inference Indexing

• US 20120158738A1
• Filed: 12/15/2010
• Published: 06/21/2012
• Est. Priority Date: 12/15/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Jiang, Li, Viola, Paul Alfred, McGovern, Andrew Paul, Ahari, Sanaz, Buehrer, Gregory T., Szymanski, Jakub Jan
• Granted Patent: US 8,977,625 B2
• Time in Patent Office: Days
• US Class Current: 707/748
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

Methods, systems, and media are provided for facilitating generation of an inference index. In embodiments, a canonical entity is referenced. The canonical entity is associated with web documents. One or more queries that, when input, result in a selection of at least one of the web documents are identified. An entity document is generated for the canonical entity. The entity document includes the identified queries and/or associated text from the content of a document or from an entity title that result in the selection of the at least one of the web documents. The entity document and corresponding canonical entity can be combined with additional related entity documents and canonical entities to generate an inference index.

Claims

1. A computer-implemented method of facilitating generation of an entity document using a computing system having processor, memory, and data storage subsystems, the computer-implemented method comprising:

referencing a canonical entity that is associated with one or more web documents;

identifying one or more queries that, when input, result in a selection of at least one of the one or more web documents; and

generating an entity document for the canonical entity, the entity document including the one or more identified queries that resulted in the selection of the at least one of the one or more web documents via the processor.

2. The computer-implemented method of claim 1, further comprising:

generating an inference index using the canonical entity and the entity document along with other related canonical entities and corresponding entity documents.

3. The computer-implemented method of claim 2, wherein the inference index corresponds with a knowledge domain of related canonical entities.

4. The computer-implemented method of claim 3, further comprising:

utilizing the inference index in response to a real-time user query to identify a particular canonical entity that is most related to the real-time user query.

5. The computer-implemented method of claim 1, wherein the entity document includes the one or more web documents.

6. The computer-implemented method of claim 1, further comprising:

composing one or more queries which are mapped to the canonical entity;

searching the composed one or more queries to obtain one or more resultant web documents mapped to the canonical entity;

collecting stored queries linked to the one or more resultant web documents; and

combining the collected stored queries to the entity document for the canonical entity.

7. The computer-implemented method of claim 1, wherein the one or more web documents associated with the canonical entity is based on a user query or a composed query.

8. The computer-implemented method of claim 1, wherein one or more uniform resource locators are associated with the canonical entity corresponding with the entity document.

9. One or more computer-readable storage media containing computer-readable instructions embodied thereon that when executed by a computing device, perform a method of utilizing an entity document, the method comprising:

receiving an input query having one or more ngrams;

accessing an inference index comprising a plurality of canonical entities and corresponding entity documents, wherein each of the entity documents comprise one or more ngrams of an input query which were derived from one or more selected web documents, the web documents related to one of the plurality of canonical entities associated with the entity document;

for a first ngram of a set of ngrams of the input query, identifying each entity document that includes a query or associated text which matches the first ngram and its corresponding score; and

identifying the canonical entity corresponding with the entity document having the highest score that matches the first ngram.

10. The computer-implemented method of claim 9 further comprising:

for a second ngram of the set of ngrams of the input query, identifying each entity document that includes a query matching the second ngram and a corresponding score of which the second ngram matches the query within the entity document.

11. The computer-implemented method of claim 10 further comprising:

determining an aggregate score for each entity document by aggregating the score of which the first ngram matches a query within the entity document and the score of which the second ngram matches a query within the entity document.

12. The computer-implemented method of claim 11 further comprising:

selecting the canonical entity associated with the entity document having the greatest aggregate score as the canonical entity relevant to the input query.

13. The computer-implemented method of claim 9, wherein the inference index is associated with a particular knowledge domain.

14. The computer-implemented method of claim 9, wherein said method is implemented at query time prior to content retrieval.

15. An inference index system having a computing device with processor, memory, and data storage subsystems, the inference index system comprising:

an inference index generating component to generate an inference index that comprises;

a plurality of canonical entities associated with a knowledge domain, and

an entity document in association with each canonical entity, the entity document comprising one or more queries or associated text that, if input, result in selection of a uniform resource locator associated with the canonical entity; and

an entity selecting component to select one of the plurality of canonical entities relevant to a search query, or a portion thereof, the canonical entity being selected from among the plurality of canonical entities within the inference index based on a cumulative score of functions associated with each of the canonical entities that match at least a portion of the search query.

16. The inference index system of claim 15, wherein the entity selecting component ranks the canonical entities based on the cumulative score associated with each of the canonical entities that match at least a portion of the search query or the associated text.

17. The inference index system of claim 16, wherein the ranked canonical entities are communicated to an index searching component for content retrieval.

18. The inference index system of claim 17, wherein the index searching component retrieves content based on the ranked canonical entities.

19. The inference index system of claim 15, wherein at least a portion of the uniform resource locators are derived from artificially composed queries.

20. The inference index system of claim 19, wherein the artificially composed queries are linked to real queries mapped to the canonical entity.