SELF-TUNING ALTERATIONS FRAMEWORK

• US 20130346400A1
• Filed: 06/20/2012
• Published: 12/26/2013
• Est. Priority Date: 06/20/2012
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: RAMSEY, WILLIAM D., CRASWELL, NICHOLAS ERIC, DUMOULIN, BENOIT
• Granted Patent: US 8,984,012 B2
• Time in Patent Office: Days
• US Class Current: 707/723
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9532 Query formulation

Abstract

Embodiment described herein are directed to an enhanced search engine with multiple feedback loops for providing optimal search results that are responsive a user'"'"'s search query. The user'"'"'s search query is parsed, and based on the underlying terms, different linguistic models and refinement techniques generate alternative candidate search queries that may yield better results. Searches are performed for the original search query and the candidate search queries, and different scores are used to select the best search results to present to the user. Results making it onto the list, as well as the underlying candidate search query, linguistic model, or refinement technique for generating that search query, will have their corresponding scores updated to reflect their success of generating a search result. Scores are stored and used by future searches to come up with better results.

Claims

1. A computer-implemented method, performed by one or more computing devices, for generating and transmitting search results responsive to search query from a user, the method comprising:

receiving the search query comprising one or more search terms;

obtaining candidate search queries for the search query;

obtaining candidate scores for the candidate search queries;

initiating a search for online resources relevant to the one or more search terms;

initiating searches for additional online resources relevant to the candidate search queries;

obtaining interaction scores for the online resources and the additional online resources;

using the candidate scores and the interaction scores, generating a search results list comprising a group of the online resources and the additional online resources;

incrementing the candidate scores of the candidate search queries that yielded the additional online resources selected for the list; and

transmitting the list of search results as a response to the search query.

2. The computer-implemented method of claim 1, wherein the candidate scores each indicate a probability that at least of the candidate search queries should replace the search query.

3. The computer-implemented method of claim 2, wherein the probability is based on historical interactions of users with results stemming from the at least one of the candidate search queries.

4. The computer-implemented method of claim 1, wherein at least one of the candidate search queries includes terms that are at least one member of a comprising a synonym, alternative spelling, corrected spelling, and homonym of at least one of the search terms.

5. The computer-implemented method of claim 1, wherein at least one of the candidate search queries includes an indication for searching a particular vertical.

6. The computer-implemented method of claim 1, wherein at least one of the candidate search queries includes an indication for restricting results to a particular geographic location.

7. The computer-implemented method of claim 1, wherein at least one of the candidate search queries includes one or more related search queries related to a context of the search query.

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

receiving an indication of an interaction of the user with at least one of the search results on the list; and

storing an indication of the user interacting with the at least one of the search results on the list.

9. The computer-implemented method of claim 8, wherein the interaction of the user comprises at least member of a group comprising a mouse click, an initiation of a hypertext transfer protocol (HTTP) request, a mouse hover, a touch gesture, an image recognition, and a hyperlink selection of at least one of the search results on the list.

10. The computer-implemented method of claim 9, wherein the mouse hover indicates a length time a mouse pointer was hovered over the at least one of the search results on the list.

11. One or more computer-storage media storing computer-executable instructions that, when executed by a processor, perform a method for generating a list of search results responsive to a search query, comprising:

monitoring candidate scores and interaction scores for a plurality of candidate search queries;

updating a candidate score of a candidate search query upon receipt of an indication a search result stemming from the candidate search query was interacted with;

monitoring interaction scores indicative of users interactions with results stemming from the plurality of candidate search queries;

selecting a group of the candidate search queries with search terms that are variants of a search query from a user; and

using the candidate scores and the interaction scores of the group of the candidate search queries to select at least one candidate search query and corresponding online resource to include in the list of search results responsive to the search query.

12. The one or more computer-storage media of claim 11, further comprising:

receiving an indication that a user interaction occurred with the corresponding online resource; and

because of the user interaction, updating an interaction score associated with the at least one candidate search query.

13. The one or more computer-storage media of claim 12, further comprising modifying a candidate score associated with the at least one search query.

14. The one or more computer-storage media of claim 13, further comprising modifying an interaction score associated with the corresponding online resource.

15. The one or more computer-storage media of claim 12, wherein the user interaction comprises at least one member of a group comprising a mouse click, an initiation of a hypertext transfer protocol (HTTP) request, a mouse hover, a hyperlink selection, a touch gesture, and an image recognition.

16. The one or more computer-storage media of claim 11, the group of then candidate search queries includes at least one member of a group comprising a synonym, alternative spelling, corrected spelling, and homonym of at least one of the search terms in the search query from the user.

17. A system, comprising:

one or more processors;

one or more computer-storage media; and

executable instructions stored on the one or more computer-storage media and executable by the one or more processors to receive a search query and generate an optimal list of search results in response, comprising;

(1) a candidate generator that determines one or more candidate search queries, each comprising variant terms to search terms of the search query,

(2) a probability generator to determine probabilities that including results stemming from searches of the one or more candidate search queries in the optimal list of results will result in a user interaction with corresponding online resources, wherein the probabilities are based on interaction scores and candidate scores for the candidate search queries, and

(3) a blending module to the search results for the optimal list based on the probabilities.

18. The system of claim 17, further comprising at least one member of a group comprising:

an error module to correct any misspellings in the search query; and

a language module to generate the variant terms.

19. The one or more computer-readable media of claim 17, further comprising:

a vertical module that generates candidate search queries refined to one or more verticals; and

a geographic module that generates candidate search queries refined to a particular geographic location.

20. The one or more computer-readable media of claim 18, further comprising a database capable of storing and updating the candidate scores and interaction scores for the candidate search queries.