Using a dialog system for integrating multiple domain learning and problem solving

1. A method, by a processor, using a dialog system for integrating multiple domain learning and problem solving, comprising:

executing machine learning logic to generate a decision model, wherein the decision model is trained, in a first training stage, to define one or more problem instances for one or more selected domains in a multi-domain database according to a problem instance template, identified user intent, links to one or more problem solvers associated with the one or more selected domains, or a combination thereof;

determining a dialog plan for the one or more problem instances using a dialog system associated with the multi-domain database, wherein each record in the multi-domain database corresponds to a selected database for the one or more selected domains, and wherein determining the dialog plan includes identifying a modeling cost-to-cognitive load sensitivity ratio of candidate solutions identified for the one or more problem instances, the modeling cost-to-cognitive load sensitivity ratio taking into consideration utility details of respective solutions of the candidate solutions correspondent to an overall cost of the respective solutions based on weights identified as preferable to the user;

providing a solution to the user for the one or more problem instances according to the decision model; and

learning one or more preferences of a user according to the solution by executing the machine learning logic to re-train, in a second training stage, the decision model according to the learned one or more preferences of the user, wherein the decision model is iteratively updated using reinforcement learning to enhance future solutions of future problem instances to provide to the user.

2. The method of claim 1 , wherein learning one or more preferences of a user according to the solution further includes executing the machine learning logic to learn preferences of one or more variables associated with a domain.

3. The method of claim 1 , wherein learning one or more preferences of a user according to the solution further includes executing the machine learning logic to learn preferences of user interface (UI) elements associated with a domain.

4. The method of claim 1 , wherein learning one or more preferences of a user according to the solution further includes executing the machine learning logic to learn preferences over a selected period of time.

5. The method of claim 1 , further including:

defining one or more variables for the one or more problem instances; or

assigning the one or more variables a predetermined value extracted from a user profile, a configuration file, and historical data associated with prior problem instances.

6. The method of claim 1 , further including:

defining one or more dependencies between the one or more variables; or

adding, deleting, or updating one or more domains of the multiple domains.

7. The method of claim 1 , further including customizing an interface of the dialog system according to the learning.

8. A system, for using a dialog system for integrating multiple domain learning and problem solving in a computing environment, comprising:

one or more processors with executable instructions that when executed cause the system to:

execute machine learning logic to generate a decision model, wherein the decision model is trained, in a first training stage, to define one or more problem instances for one or more selected domains in a multi-domain database according to a problem instance template, identified user intent, links to one or more problem solvers associated with the one or more selected domains, or a combination thereof;

determine a dialog plan for the one or more problem instances using a dialog system associated with the multi-domain database, wherein each record in the multi-domain database corresponds to a selected database for the one or more selected domains, and wherein determining the dialog plan includes identifying a modeling cost-to-cognitive load sensitivity ratio of candidate solutions identified for the one or more problem instances, the modeling cost-to-cognitive load sensitivity ratio taking into consideration utility details of respective solutions of the candidate solutions correspondent to an overall cost of the respective solutions based on weights identified as preferable to the user;

provide a solution to the user for the one or more problem instances; and

learn one or more preferences of a user according to the solution.

9. The system of claim 8 , wherein learning one or more preferences of a user according to the solution further executes the machine learning logic to learn preferences of one or more variables associated with a domain.

10. The system of claim 8 , wherein learning one or more preferences of a user according to the solution further executes the machine learning logic to learn preferences of user interface (UI) elements associated with a domain.

11. The system of claim 8 , wherein learning one or more preferences of a user according to the solution further executes the machine learning logic to learn preferences over a selected period of time.

12. The system of claim 11 , wherein the executable instructions further:

define one or more variables for the one or more problem instances; or

assign the one or more variables a predetermined value extracted from a user profile, a configuration file, and historical data associated with prior problem instances.

13. The system of claim 11 , wherein the executable instructions further:

define one or more dependencies between the one or more variables; or

add, delete, or update one or more domains of the multiple domains.

14. The system of claim 11 , wherein the executable instructions further customize an interface of the dialog system according to the learning.

15. A computer program product for, by one or more processors, using a dialog system for integrating multiple domain learning and problem solving in a computing environment, the computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:

an executable portion that executes machine learning logic to generate a decision model, wherein the decision model is trained, in a first training stage, to define one or more problem instances for one or more selected domains in a multi-domain database according to a problem instance template, identified user intent, links to one or more problem solvers associated with the one or more selected domains, or a combination thereof;

an executable portion that determines a dialog plan for the one or more problem instances using a dialog system associated with the multi-domain database, wherein each record in the multi-domain database corresponds to a selected database for the one or more selected domains, and wherein determining the dialog plan includes identifying a modeling cost-to-cognitive load sensitivity ratio of candidate solutions identified for the one or more problem instances, the modeling cost-to-cognitive load sensitivity ratio taking into consideration utility details of respective solutions of the candidate solutions correspondent to an overall cost of the respective solutions based on weights identified as preferable to the user;

an executable portion that provides a solution to the user for the one or more problem instances; and

an executable portion that learns one or more preferences of a user according to the solutions.

16. The computer program product of claim 15 , wherein the executable portion that learns the one or more preferences of the user according to the solution further executes the machine learning logic to learn preferences of one or more variables associated with a domain.

17. The computer program product of claim 15 , wherein the executable portion that learns the one or more preferences of the user according to the solution further executes the machine learning logic to learn preferences of user interface (UI) elements associated with a domain.

18. The computer program product of claim 15 , wherein the executable portion that learns the one or more preferences of the user according to the solution further executes that machine learning logic to learn preferences over a selected period of time.

19. The computer program product of claim 15 , further including an executable that:

defines one or more variables for the one or more problem instances;

assigns the one or more variables a predetermined value extracted from a user profile, a configuration file, and historical data associated with prior problem instances;

defines one or more dependencies between the one or more variables; or

adds, deletes, or updates one or more domains of the multiple domains.

20. The computer program product of claim 15 , further including an executable that customizes an interface of the dialog system according to the learning.