Method for collaborative knowledge base development
A knowledge base is collaboratively developed by receiving language input from at least one editor, featurizing it into language elements, extracting predicate sets that are missing predicate arguments, querying editors for input regarding the missing arguments, and updating the knowledge base with suitable predicate arguments and other language elements provided by the editors.
1. An automated method for collaboratively developing a knowledge base with information from a plurality of editors comprising the steps of:
receiving, from at least one editor, a language input;
featurizing said language input to produce a featurized data;
extracting, based upon said featurized data, an incomplete predicate set comprising a predicate head;
determining whether a predicate argument is present in said incomplete predicate set; and
querying at least one of the plurality of editors for an input regarding said predicate argument.
2. The method of claim 1 , wherein the step of querying comprises the step of communicating to the at least one of the plurality of editors a natural language question regarding said predicate argument.
3. The method of claim 1 ,
wherein the step of querying comprises the step of posting a query to an electronic bulletin board; and
further comprising the step of requesting the at least one of the plurality of editors to provide input regarding said predicate argument.
4. The method of claim 1 , further comprising the steps of:
receiving an input predicate argument from at least one of the plurality of editors;
pairing said input predicate argument with said predicate head to generate a complete predicate set; and
storing said complete predicate set in said knowledge base.
5. The method of claim 1 , further comprising the steps of:
receiving an input predicate argument from at least one of the plurality of editors; and
storing said featurized data and said input predicate argument as a plurality of exemplary language features associated with said incomplete predicate set.
6. The method of claim 1 , wherein the step of determining comprises the step of determining whether a complete version of said incomplete predicate set is already present in a use case knowledge base; and retrieving said complete version from said use case knowledge base.
7. The method of claim 1 , wherein the plurality of editors each have a designated authority level determining a permission level for providing said input regarding said predicate argument.
8. The method of claim 1 , further comprising the steps of:
receiving an input predicate argument from at least one of the plurality of editors;
determining if a predicate set confidence level is above a threshold;
pairing said input predicate argument with said predicate head to generate a complete predicate set; and
storing said complete predicate set in said knowledge base.
9. The method of claim 1 , further comprising the steps of:
receiving an input predicate argument from at least one of the plurality of editors;
determining if a predicate set confidence level is above a threshold;
determining whether a predicate argument information has been provided in said featurized data; and
storing said predicate argument information and said featurized data as exemplary language features associated with a complete predicate set.
10. An automated method for collaboratively developing a knowledge base with information from a plurality of editors comprising the steps of:
designating a plurality of editors;
providing to at least one of the plurality of editors a knowledge base activity information comprising an unfulfilled predicate argument information associated with an unfulfilled predicate argument;
inviting an editor input from the at least one of the plurality of editors regarding said unfulfilled predicate argument information;
fulfilling said unfulfilled predicate argument information based upon said editor input to generate a complete predicate set; and
storing said complete predicate set in said knowledge base.
11. The method of claim 10 , wherein the step of providing to at least one of the plurality of editors knowledge base activity information comprises the step of posting said knowledge base activity information on an electronic bulletin board accessible to at least one of the plurality of editors.
12. The method of claim 10 , wherein the step of inviting editor input comprises the step of providing a natural language query relating to said unfulfilled predicate argument.
13. The method of claim 10 , wherein the plurality of editors each have a designated authority level determining a permission level for providing said editor input regarding said unfulfilled predicate argument information.
14. The method of claim 10 , further comprising the steps of:
receiving an input predicate argument from at least one of the plurality of editors;
pairing said input predicate argument with said predicate head to generate a complete predicate set; and
storing said complete predicate set in said knowledge base.
15. The method of claim 10 , further comprising the steps of:
receiving an input predicate argument from at least one of the plurality of editors; and
storing said input predicate argument as exemplary language features associated with said unfulfilled predicate argument.