Recommending solutions with an expert system

• US 20070143239A1
• Filed: 12/07/2006
• Published: 06/21/2007
• Est. Priority Date: 12/20/2005
• Status: Active Grant

• Current Assignee: Twitter Inc (X Holdings Corp.)
• Original Assignee: International Business Machines Corporation
• Inventors: Boyle, Peter, MacAulay, Alexander
• Granted Patent: US 7,702,601 B2
• Time in Patent Office: Days
• US Class Current: 706/45
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

An expert system is used to recommend a customized solution for a customer. The expert system includes a set of scenarios each of which includes default facts and is associated with a detailed rule base that when applied in isolation to the default facts fully determines a prototype solution. A particular scenario is selected from the set based at least on user input. A customized solution is generated by applying the detailed rule base associated with the particular scenario to facts based at least on the user input and any default facts of the particular scenario that complement and do not conflict with the facts that are based at least on the user input. The customized solution is iteratively refined.

Claims

1. A method for using an expert system to recommend a customized solution for a customer, the method comprising:

providing in the expert system a set of scenarios each of which includes default facts and is associated with a detailed rule base that when applied in isolation to the default facts fully determines a prototype solution;

selecting a particular scenario from the set of scenarios based at least on user input;

generating a customized solution by applying the detailed rule base associated with the particular scenario to facts based at least on the user input and any default facts of the particular scenario that complement and do not conflict with the facts based at least on the user input; and

iteratively refining the customized solution.

2. The method of claim 1, wherein selecting a particular scenario from the set of scenarios based at least on user input comprises:

posing at least one specific question to a user who is a representative of the customer and obtaining supplied facts from answers to the questions;

in a context of the supplied facts, inferring facts from parseable input supplied to the expert system;

determining a subset of the scenarios from the inferred facts and the supplied facts; and

enabling the user to select a particular scenario from the subset of the scenarios.

3. The method of claim 2, further comprising:

enabling the user to edit the parseable input; and

revising the inferred facts according to the edited parseable input in the context of the supplied facts.

4. The method of claim 2, wherein generating the customized solution comprises:

generating the customized solution by applying the detailed rule base associated with the particular scenario to a set of candidate facts that comprises the inferred facts, the supplied facts, and assumed facts, where the assumed facts are any default facts of the particular scenario that complement and do not conflict with the inferred facts and the supplied facts.

5. The method of claim 4, further comprising:

presenting the user with a log of the parseable input; and

indicating to the user which of the candidate facts are most significant in determining the customized solution.

6. The method of claim 5, wherein iteratively refining the customized solution comprises:

if the user challenges any of the candidate facts, updating the candidate facts by replacing each challenged fact with at least one supplied fact from answers to specific questions posed to the user; and

refining the customized solution by applying the detailed rule base associated with the particular scenario to the updated candidate facts.

7. The method of claim 5, wherein iteratively refining the customized solution comprises:

enabling the user to edit the log;

updating the inferred facts and the assumed facts of the candidate facts in the context of the supplied facts and the edited log; and

refining the customized solution by applying the detailed rule base associated with the particular scenario to the updated candidate facts.

8. The method of claim 4, further comprising:

if the refined customized solution is underdetermined, posing additional specific questions to obtain additional supplied facts.

9. The method of claim 4, further comprising:

if the refined customized solution is underdetermined, adding additional assumed facts to the set of candidate facts.

10. The method of claim 4, further comprising:

if iteratively refining the customized solution does not converge to a final solution, obtaining input from a subject matter expert.

11. The method of claim 10, further comprising:

revising the expert system with the input from the subject matter expert.

12. The method of claim 11, further comprising:

inviting the user to resume iteratively refining the customized solution using the revised expert system.

13. A computer program product comprising a computer useable medium having computer readable executable modules comprising:

a set of scenarios each of which includes default facts and is associated with a detailed rule base that when applied in isolation to the default facts fully determines a prototype solution;

a scenario selection module to select a particular scenario from the set of scenarios based at least on user input, using a scenario selection rule base; and

a solution refinement module to generate a customized solution by applying the detailed rule base associated with the particular scenario to facts based at least on the user input and any default facts of the particular scenario that complement and do not conflict with the facts based at least on the user input, and to iteratively refine the customized solution.

14. The computer program product of claim 13, wherein the scenario selection module includes a decision tree reasoning mechanism to use the scenario selection rule base.

15. The computer program product of claim 13, wherein the solution refinement module uses a backward chaining rules engine.

16. The computer program product of claim 13, wherein the scenario selection rule base comprises specific questions to be posed to a user, and wherein supplied facts are obtained from answers to the specific questions.

17. The computer program product of claim 16, wherein the computer useable medium has further computer readable executable modules comprising:

a parsing module to parse parseable input in the context of the supplied facts and infer therefrom inferred facts.

18. An expert system for recommending a customized solution for a customer, the expert system comprising:

a set of scenarios each of which includes default facts and is associated with a detailed rule base that when applied in isolation to the default facts fully determines a prototype solution;

a system for selecting a particular scenario from the set of scenarios based at least on user input;

a system for generating a customized solution by applying the detailed rule base associated with the particular scenario to facts based at least on the user input and any default facts of the particular scenario that complement and do not conflict with the facts based at least on the user input; and

a system for iteratively refining the customized solution.

19. A method for deploying an application for recommending a customized solution, using an expert system, for a customer, comprising:

providing a computer infrastructure being operable to;

provide in the expert system a set of scenarios each of which includes default facts and is associated with a detailed rule base that when applied in isolation to the default facts fully determines a prototype solution;

select a particular scenario from the set of scenarios based at least on user input;

generate a customized solution by applying the detailed rule base associated with the particular scenario to facts based at least on the user input and any default facts of the particular scenario that complement and do not conflict with the facts based at least on the user input; and

iteratively refine the customized solution.