Machine learning query session enhancement

Abstract

A computer method, system, and device of training an empathy model for detecting a type of query, the method including defining an intent for detecting closed ended queries, providing a plurality of queries that are closed ended queries to a machine learning model generator, said plurality of queries comprising training data, providing a plurality of corresponding labels identifying the plurality of queries as closed ended queries, and generating a model that classifies closed ended queries as a function of the training data.

1 . A computer implemented method of generating query suggestions, the method comprising:
 receiving queries as the queries are posed;   receiving indicia, from a language understanding machine learning model, representative of a received query being closed ended; and   generating a new query that is more open-ended than the query associated with the received indicia.   
     
     
         2 . The method of  claim 1  wherein the new query is selected from a set of queries in a table. 
     
     
         3 . The method of  claim 2  wherein the new query is generated to elicit more information than information received responsive to the query associated with the received indicia. 
     
     
         4 . The method of  claim 3  wherein the new query is generated as a function of received responses to the query associated with the received indicia, wherein the new query is provided to a user as a next query in the received queries. 
     
     
         5 . The method of  claim 1  wherein the language understanding machine learning model trained with a plurality of queries that are labeled closed ended queries, and wherein the indicia includes a confidence level value, and wherein the indicia includes a confidence level value. 
     
     
         6 . The method of  claim 1  wherein generating a new query is performed by a whisper bot. 
     
     
         7 . A computer implemented method of generating query suggestions, the method comprising:
 receiving queries as the queries are posed;   receiving indicia, from a language understanding model, representative of an intent of a received query; and   generating a new communication in response to the indicia representative of the intent of the received query.   
     
     
         8 . The method of  claim 6  wherein the new communication is selected from a set of communications in a table. 
     
     
         9 . The method of  claim 8  wherein the new communication is generated to elicit more information than information received responsive to the query associated with the received indicia. 
     
     
         10 . The method of  claim 9  wherein the new communication is generated as a function of received responses to the query associated with the received indicia, wherein the new communication is provided to a user as a next query in the received queries. 
     
     
         11 . The method of  claim 7  wherein the language understanding machine learning model is trained with a vivid grammar and a domain specific grammar. 
     
     
         12 . The method of  claim 11  wherein the indicia includes a confidence level value. 
     
     
         13 . The method of  claim 7  wherein the indicia includes data representative of at least one of a communication confirmation, a job to be done, a leading question, a problem, and a yes/no question. 
     
     
         14 . The method of  claim 13  wherein the data corresponding to a job to be done or a problem includes at least one indication of how, how much, where, who/what, and why. 
     
     
         15 . A device comprising:
 a processor; and   a memory device coupled to the processor and having a program stored thereon for execution by the processor to perform operations comprising:
 receiving queries as the queries are posed; 
 receiving indicia, from a language understanding model, representative of an intent of a received query; and 
 generating a new communication in response to the indicia representative of the intent of the received query. 
   
     
     
         16 . The device of  claim 16  wherein the new communication is selected from a set of communications in a table. 
     
     
         17 . The device of  claim 16  wherein the new communication is generated to elicit more information than information received responsive to the query associated with the received indicia. 
     
     
         18 . The device of  claim 17  wherein the new communication is generated as a function of received responses to the query associated with the received indicia, wherein the new communication is provided to a user as a next query in the received queries. 
     
     
         19 . The device of  claim 15  wherein the language understanding machine learning model is trained with a vivid grammar and a domain specific grammar, wherein the indicia includes a confidence level value, and wherein the indicia includes data representative of at least one of a communication confirmation, a job to be done, a leading question, a problem, and a yes/no question. 
     
     
         20 . The device of  claim 15  wherein the data corresponding to a job to be done or a problem includes at least one indication of how, how much, where, who/what, and why.