Method for collaborative knowledge base development
A use case knowledge base is collaboratively developed by receiving language input from a user, featurizing it into language elements, extracting predicate sets that are missing a predicate head or a predicate argument, querying users for input regarding the missing predicate information, and updating the knowledge base with predicate sets and other information provided by the users.
1 . An automated method for developing a computer knowledge base storage with information from a plurality of sources comprising the steps of: receiving, from at least one source, a language input; vectorizing said language input, using at least one language featurization system trained on a reference speech data comprising featurized audio from clinical trial data measured on a clinically recognized evaluation instrument, to produce vectorized data comprising a plurality of vectors; extracting, based upon said plurality of vectors, an incomplete predicate set comprising a predicate head without a predicate argument; determining a confidence level for said incomplete predicate set; comparing said confidence level to a predetermined confidence threshold; and querying at least one source for an input regarding at least one predicate argument to create a complete predicate set.
2 . The automated method of claim 1 ,
wherein said input comprises an input predicate argument; and
the automated method further comprises the steps of:
pairing said input predicate argument with said predicate head to generate a complete predicate set; and
storing said complete predicate set in said computer knowledge base storage.
3 . The automated method of claim 2 , wherein said reference speech data comprises featurized audio from clinical trial data measured on a clinically recognized Generalized Anxiety Disorder (GAD) evaluation instrument.
4 . The automated method of claim 2 , wherein said reference speech data comprises featurized audio from clinical trial data measured on a clinically recognized quality of life (QoL) evaluation instrument.
5 . The automated method of claim 2 , further comprising the step of determining whether said incomplete predicate set is already in said computer knowledge base storage.
6 . The automated method of claim 2 , wherein the step of pairing comprises the step of determining whether said predicate head and said input predicate argument are sufficient to generate a complete predicate set.
7 . The automated method of claim 1 wherein the step of querying at least one source comprises the step of querying at least one editor for said input regarding the at least one predicate argument.
8 . The automated method of claim 7 , wherein the step of querying at least one editor comprises the step of communicating to the at least one editor a natural language question regarding the at least one predicate argument.
9 . The automated method of claim 7 , wherein the step of querying at least one editor comprises the steps of: formulating a query regarding the at least one predicate argument; and posting said query to an electronic bulletin board.
10 . The automated method of claim 9 , further comprising the steps of: receiving an input predicate argument from the at least one editor; pairing said predicate head with said input predicate argument to generate a complete predicate set; and storing said complete predicate set in said computer knowledge base storage.
11 . The automated method of claim 10 , further comprising the step of:
storing said featurized data with said complete predicate set in said computer knowledge base storage.
12 . The automated method of claim 1 wherein the step of querying at least one source comprises the step of querying at least one electronic source for at least one input regarding said predicate argument.
13 . The automated method of claim 12 wherein the step of querying at least one electronic source comprises the step of posting a query for said input regarding the at least one predicate argument to an electronic bulletin board.
14 . An automated method for developing a computer knowledge base storage with information from a plurality of sources comprising the steps of:
receiving, from at least one source, a language input;
vectorizing said language input, using at least one language featurization system trained on a reference speech data comprising featurized audio from clinical trial data measured on a clinically recognized evaluation instrument, to produce a vectorized data comprising a plurality of vectors;
extracting, based upon said plurality of vectors, an incomplete predicate set comprising a predicate head without a predicate argument;
determining a confidence level for said incomplete predicate set;
comparing said confidence level to a predetermined confidence threshold;
querying at least one source for an input regarding said incomplete predicate set;
determining a second confidence level for said incomplete predicate set based on said input regarding said incomplete predicate set; and
comparing said second confidence level to said predetermined confidence threshold.
15 . The automated method of claim 14 wherein the step of querying at least one source comprises the step of querying at least one editor for said input regarding said incomplete predicate set.
16 . The automated method of claim 15 , wherein the step of querying at least one editor comprises the step of communicating to the at least one editor a natural language question regarding said incomplete predicate set.
17 . The automated method of claim 15 , wherein the step of querying at least one editor comprises the steps of:
formulating a query regarding said incomplete predicate set; and
posting said query to an electronic bulletin board.
18 . The automated method of claim 14 wherein the step of querying at least one source comprises the step of querying at least one electronic source for said input regarding said incomplete predicate set.
19 . The automated method of claim 18 wherein the step of querying at least one electronic source comprises the step of posting a query for said input regarding said incomplete predicate set to an electronic bulletin board.