Knowledge base comprising executable stories

• US 20060112045A1
• Filed: 10/05/2004
• Published: 05/25/2006
• Est. Priority Date: 10/05/2004
• Status: Abandoned Application

• Current Assignee: Northrop Grumman Corporation
• Original Assignee: Northrop Grumman Corporation
• Inventors: Ellis, Dennis Regan, Talbot, Patrick James, McCartney, Jeffrey Allen
• Application Number:
  US10/958,753
• Publication Number:
  US 20060112045A1
• Time in Patent Office: Days
• US Class Current: 706/46
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

Systems and methods are provided for assisted decision-making. A knowledge base comprises a plurality of executable stories. A given story comprises a belief network containing at least one associated hypothesis and evidence associated with the at least one hypothesis. The system further comprises at least one decision algorithm having respective associated input formats. A translation interface is operative to convert a given story from the knowledge base into a decision network having a format associated with the at least one decision algorithm.

Claims

1. An assisted decision-making system comprising:

a knowledge base comprising a plurality of executable stories, a given executable story comprising a belief network containing at least one associated hypothesis and evidence associated with the at least one hypothesis;

at least one decision algorithm having respective associated input formats; and

a translation interface operative to convert a given story from the knowledge base into a decision network having a format associated with the at least one decision algorithm.

2. The system of claim 1, at least one of the plurality of executable stories comprising a Dempster-Shafer belief network.

3. The system of claim 1, the system further comprising a fusion engine that is operative to add, remove, and modify the at least one hypothesis associated with a given story.

4. The system of claim 3, the fusion engine being operative to determine the logical intersection between a first belief network, associated with a first story, and a second belief network, associated with a second story.

5. The system of claim 3, the fusion engine being operative to determine the logical union between a first belief network, associated with a first story, and a second belief network, associated with a second story.

6. The system of claim 1, the translation interface comprising first and second interface modules, associated respectively with first and second data processing algorithms, each of the first and second interface modules being operative to convert a story from the knowledge base into the appropriate input format for its associated data processing algorithm.

7. The system of claim 1, further comprising at least one information extractor that reduces evidence from an external source into an evidence template that is appropriately formatted for the knowledge base.

8. The system of claim 7, further comprising at least one evidence classifier that evaluates the evidence template to associate the template with a hypothesis within the knowledge base.

9. The system of claim 7, the at least one decision algorithm comprising a data mining algorithm that examines at least one decision network provided through the translation interface to determine one or more appropriate hypotheses to associate with the evidence template.

10. The system of claim 1, the at least one decision algorithm comprising a case based reasoning system that modifies the decision network according to an analysis of at least one executable story provided to the case based reasoning algorithm from the knowledge base through the translation interface.

11. A method for assisting decision making comprising:

storing a plurality of executable stories in a knowledge base;

converting a given story from the plurality of executable stories into a decision network having a format associated with a decision algorithm associated with the knowledge base;

processing the decision network at the decision algorithm to produce a revised decision network; and

converting the revised decision network into a revised executable story.

12. The method of claim 11, further comprising fusing a first executable story and a second executable story from the plurality of executable stories to create a third executable story.

13. The method of claim 11, further comprising reducing evidence from an external source into an evidence template that is appropriately formatted for the knowledge base.

14. The method of claim 13, further comprising evaluating a given evidence template to associated the template with a hypothesis within the knowledge base.

15. The method of claim 11, wherein the format associated with the decision network is a directed graph.

16. The method of claim 11, wherein the format associated with the decision network comprises tabular data.

17. An assisted decision-making system comprising:

means for storing a plurality of executable stories, a given story comprising a belief network containing at least one associated hypothesis and evidence associated with the at least one hypothesis;

means for processing a decision network, the means for processing having an associated input format; and

means for translating a given story from the means for storing into a decision network having the input format associated with the means for processing.

18. The system of claim 17, further comprising means for ranking stories in a knowledge base.

19. The system of claim 18, the means for ranking including means for ranking stories according to an average belief in the hypotheses that constitute the underlying belief network.

20. The system of claim 18, the means for ranking including means for ranking stories according to an average belief in the hypotheses that constitute the underlying belief network.

21. The system of claim 17, the means for translating including means for translating a decision network from the input format associated with the means for processing into an executable story.

22. The system of claim 17, further comprising means for fusing a first executable story and a second executable story from the plurality of executable stories to create a third executable story.

23. The system of claim 17, further comprising means for reducing evidence from an external source into an evidence template that is appropriately formatted for the means for storing.