Question answering over structured content on the web

• US 20070094285A1
• Filed: 10/21/2005
• Published: 04/26/2007
• Est. Priority Date: 10/21/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Burges, Christopher, Brill, Eric, Agichtein, Yevgeny
• Granted Patent: US 7,873,624 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2428 Query predicate definition using graphical user interfaces, including menus and forms G06F16/2423 takes precedence
• CPC Class Codes: G06F 16/243 Natural language query formulation
• CPC Class Codes: G06F 16/24522 Translation of natural language queries to structured queries

Abstract

Structured content and associated metadata from the Web are leveraged to provide specific answer string responses to user questions. The structured content can also be indexed at crawl-time to facilitate searching of the content at search-time. Ranking techniques can also be employed to facilitate in providing an optimum answer string and/or a top K list of answer strings for a query. Ranking can be based on trainable algorithms that utilize feature vectors for candidate answer strings. In one instance, at crawl-time, structured content is indexed and automatically associated with metadata relating to the structured content and the source web page. At search-time, candidate indexed structured content is then utilized to extract an appropriate answer string in response to a user query.

Claims

1. A system that facilitates determination of search query answers, comprising:

an extraction component that obtains structured content and/or implicitly structured content from the Internet along with associated metadata; and

a query response component that employs, at least in part, the structured content to facilitate in determining at least one query answer string.

2. The system of claim 1 further comprising:

an indexing component that indexes the structured content from the extraction component utilizing, at least in part, the associated metadata;

the indexed structured content employed by the query response component to facilitate in determining the query answer string.

3. The system of claim 1, the query response component employs, at least in part, personalized data and/or localized data to facilitate in determining the query answer string.

4. The system of claim 3, the query response component reads and/or stores the personalized data on at least one remote and/or local central data server.

5. The system of claim 1, the query response component provides a top K number of query answer strings to a user based on, at least in part, similarity of the metadata, where K is an integer from one to infinity.

6. The system of claim 5, the query response component automatically orders the top K query answer strings to the user based on, at least in part, a ranking process.

7. The system of claim 1, the query response component represents the candidate answer strings as feature vectors representative of characteristics of a desired query answer string.

8. The system of claim 1 further comprising:

a ranking component that ranks answer strings to facilitate the query response component in determining at least one most likely user desirable query answer string.

9. The system of claim 8, the automatically trained ranking component employs training data to facilitate in determining query answer string rank.

10. The system of claim 9, the training data comprising, at least in part, user provided data and/or feature data associated with at least one correct desired query answer string, obtained for previous queries.

11. The system of claim 1, the metadata comprising, at least in part, network related information that is associated with the structured content.

12. The system of claim 1, the metadata comprising, at least in part, page and/or site-related information that is associated with the structured content.

13. A method for facilitating determination of search query answers, comprising:

extracting structured content and associated metadata from the Internet; and

determining at least one answer string for a query via employment, at least in part, of the structured content.

14. The method of claim 13 further comprising:

indexing the structured content utilizing, at least in part, the associated metadata; and

determining the answer string for the query based on the indexed structured content.

15. The method of claim 13 further comprising:

determining candidate query answer strings from the indexed structured content;

creating at least one feature vector representative of a desired query answer string; and

ranking the candidate query answer strings based on a set of features and the feature vector to provide a top K list of query answer strings.

16. The method of claim 15 further comprising:

employing a trainable ranking process to facilitate ranking of the candidate query answer strings.

17. The method of claim 16 further comprising:

utilizing personalized and/or localized data to facilitate in training the ranking process.

18. The method of claim 15 further comprising:

employing a probabilistic ranking model to facilitate in determining the order of the candidate query answer strings.

20. A device employing the method of claim 13 comprising at least one selected from the group consisting of a computer, a server, and a portable electronic device.

19. A system that facilitates determination of search query answers, comprising:

means for extracting structured content from the Internet along with associated metadata;

means for indexing the structured content utilizing, at least in part, the associated metadata; and

means for determining at least one answer string for a query via employment, at least in part, of the indexed structured content.