METHOD AND SYSTEM OF MANAGEMENT OF QUERIES FOR CROWD SEARCHING

• US 20140019435A1
• Filed: 07/16/2012
• Published: 01/16/2014
• Est. Priority Date: 07/16/2012
• Status: Active Grant

• Current Assignee: Politecnico di Milano
• Original Assignee: Politecnico di Milano
• Inventors: CERI, Stefano, Brambilla, Marco, Bozzon, Alessandro
• Granted Patent: US 8,825,701 B2
• Time in Patent Office: Days
• US Class Current: 707/712
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9536 Search customisation based on social or collaborative filtering
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

A computer-implemented method of management of queries for crowd searching is presented. In the method, an input model including input data e structured queries using query operators is mapped into an output model, which is obtained by modifying the input data and by adding the answers to the structured queries. Input data comprise at least one search object, which may be imported from a search system, at least one structured query operator corresponding to social interactions and a human-generated question related to the at least one search object. Mapping of the input model is according to a query task execution plan that defines a query task to be solved by a crowd, wherein mapping comprises selecting at least one social search engine running on a respective social platform and selecting one or more groups of responders interacting with the at least one social search engines and targeted to respond to the query task.

Claims

1. A computer-implemented method of management of queries for crowd searching, the method comprising:

providing an input template to a master user for coding a query for crowd searching, the input template having an input format;

receiving input data in the input template according to the input format, the input data comprising at least one search object, at least one structured query operator corresponding to social interactions and a human-generated question related to the at least one search object;

assembling the input data in the filled-in input template to generate an input model constructing a query task for crowd searching;

processing the input model according to a query task execution plan that defines the query task to be solved by a crowd, wherein processing comprises selecting at least one social search engine running on a respective social platform and selecting one or more groups of responders interacting with the at least one social search engines and targeted to respond to the query task;

instantiating, based on the query task execution plan, a plurality of platform-aware query templates, each query template having a data structure supported by the at least one social search engine;

importing the plurality of platform-aware query templates into the respective social platform through a plurality of respective responder application user interfaces allowing the one or more groups of responders to enter an answer to the query;

collecting the answers through interaction with responders of the one or more groups of responders;

processing the collected answers so as to generate an output model instantiated by an output template, and

rendering the output template as a query result to the master user.

2. The method of claim 1, wherein providing the input template to a master user is through a master application user interface and comprises prompting the master user to select at least one structured operator from a plurality of structured query operators available in the input template through the master application user interface.

3. The method of claim 1, wherein:

the input template comprises a layout of data fields for the incoming input data, and

receiving input data comprises;

inserting, by the master user, the human-generated question on a first data field of the input template,

selecting, by the master user, at least structured one query operator from a plurality of structured query operators comprised in the input template as entries of at least one second data field, and

importing the at least one search object into the input template through a search interface of a search system.

4. The method of claim 1, wherein rendering the output template comprises exporting the output template to a search interface of a search system, the method further comprising, after rendering the output template, performing at least one of the following actions:

running a search process within the search system, the search process being at least partly based on the collected answers, and

repeating providing an input template to the master user and receiving input data, wherein the input data comprise at least part of the collected answers.

5. The method of claim 1, wherein collecting answers comprises collecting by means of a plurality of platform-aware templates which are produced automatically as result of the query task execution plan.

6. The method of claim 1, wherein processing the collected answers comprises assembling the collected answers and rendering the output template to the master user comprises rendering the output template through a master application user interface by filling the output template with the collected answers.

7. The method of claim 1, wherein the at least one search object is a plurality of search objects and the input model is a tuple <

C, N, S>

, where C is a data collection comprising the plurality of search objects, N is at least one natural language question, and S is at least one structured query operator.

8. The method of claim 7, wherein the data collection C is one or more tuples described by a relational schema instance.

9. The method of claim 7, wherein the output model is a tuple <

C′

, S′

>

obtained by processing the input model according to the query task execution plan, with C′

being a data collection comprising a plurality of search objects and S′

is at least one structured answer to the at least one query operator S.

10. The method of claim 7, wherein processing the input model according to a query task execution plan is defined by a mapping model to be applied to the input model, the mapping model being a tuple comprising <

E, G, D>

, with E being the at least one social search engine, G being the one or more groups of responders interacting with the at least one social search engine and D the plurality of platform-aware query templates.

11. The method of claim 1, wherein the responder user interface is embedded in a embedding environment of the social platform by using a plug-in program logic in conjunction with the at least one social search engine.

12. The method of claim 11, wherein embedding comprises loading each responder application user interface of the plurality of responder application user interfaces on a client node of a responder of the one or more groups of responders.

13. The method of claim 1, wherein the at least one social search engine comprises at least one native widget running on a graphical user interface and the plurality of responder application user interfaces is rendered to the one or more groups of responders by using the at least one native widget of the social platform.

14. The method of claim 13, wherein the at least one social search engine comprises a plurality of application programming interfaces allowing access to the at least one native widget and to a database within the social platform for selecting the one or more groups of responders.

15. The method of claim 1, wherein importing the plurality of platform-aware query templates comprises rendering each platform-aware query template to a respective responder of the one or more groups of responders through a respective responder application user interface.

16. The method of claim 1, wherein the at least one structured query operator is one or more operators selected from the group consisting of at least one preference query operator, at least one rank query operator, at least one cluster query operator and at least one modification query operator.

17. The method of claim 16, wherein the at least one preference query operator is at least one of the operators like query, dislike query, recommend query, and tag query.

18. The method of claim 16, wherein the at least one rank query operator is at least one of the operators order query and score query.

19. The method of claim 16, wherein the at least one cluster query operator is at least one of the operators group query, order group query, merge group query, and top group query.

20. The method of claim 16, wherein the at least one modification query operator is at least one of the operators insert query, delete query, correct query, and connect query.

21. The method of claim 1, wherein processing according to a query task execution plan further comprises setting a time threshold for the collection of answers.

22. A computer-readable medium encoded with computer-executable instructions for performing the method of claim 1.