OPEN-WORLD MODELING

• US 20090006297A1
• Filed: 06/28/2007
• Published: 01/01/2009
• Est. Priority Date: 06/28/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Horvitz, Eric J., Krumm, John C., Subramani, Murugesan S.
• Granted Patent: US 7,991,718 B2
• Time in Patent Office: Days
• US Class Current: 706/46
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

The claimed subject matter provides systems and/or methods that facilitate generating an inference about events that may not have yet been observed. Open-world modeling can be used to take a history of observation so as to understand trends over time in the revelation of previously unseen events, and to make inferences with subsets of data that new unseen events will be seen. Thus, inaccuracies associated with predictions generated from incomplete data sets can be mitigated. To yield such predictions, open-world submodels and closed-world submodels that do not allow for previously unseen events can be combined via a model mixture methodology, which fuses inferences from the open- and close-world models.

Claims

1. A system that facilitates generating an inference, comprising:

a component that captures an evolution of observations over time; and

a modeling component that generates predictions of events based at least in part upon an open-world model that provides probabilities of events that have not been observed.

2. The system of claim 1, the modeling component reasons about a likelihood of seeing previously unobserved states given a specific observational context.

3. The system of claim 1, the modeling component effectuates open-world decomposition to generate the predictions.

4. The system of claim 1, the modeling component further comprising a closed-world submodeling component that performs inferences based on the observations.

5. The system of claim 1, the modeling component further comprising an open-world submodeling component that explicitly considers a likelihood of observing the non-observed events over a future period of time.

6. The system of claim 5, the open-world submodeling component models the probabilities related to the non-observed events based upon a spatial distribution.

7. The system of claim 6, the open-world submodeling component models the spatial distribution as a probability distribution over a distance from previously observed data.

8. The system of claim 1, the modeling component further comprising a weighting component that combines an open-world assumption and a closed-world assumption based upon a probability associated with at least one of the open-world assumption and the closed-world assumption being correct.

9. The system of claim 8, the weighting component fuses a weighted combination of the open-world assumption and the closed-world assumption as a function of duration of data collection.

10. The system of claim 9, the weighting component employs curve fitting to model a change in the likelihood that a user will visit a new location over time.

11. The system of claim 8, further comprising a profile component that generates a weight utilized by the weighting component based upon a consideration of user related characteristics.

12. The system of claim 1, further comprising a content component that provides relevant information associated with the generated predictions.

13. A method that facilitates handling incompleteness of data sets to enable generating a prediction, comprising:

receiving observed data;

extrapolating the observed data to yield probabilities related to non-observed data; and

generating a prediction based at least in part upon the probabilities related to the non-observed data.

14. The method of claim 13, extrapolating the observed data further comprises performing trend analysis to characterize how observational data changes given ongoing data collection.

15. The method of claim 13, generating the prediction further comprising:

generating a closed-world submodel based upon the observed data;

generating an open-world submodel by evaluating a likelihood of observing the non-observed data over a future period of time; and

combining the closed-world submodel and the open-world submodel based upon a weight to yield the prediction.

16. The method of claim 15, the weight varies dynamically with a length of observation.

17. The method of claim 15, combining the closed-world submodel and the open-world submodel further comprising:

determining characteristics of a user;

identifying a subgroup to which the user belongs based upon the characteristics;

forecasting a trajectory of a submodel weight that varies as a function of observation time based upon the identified subgroup; and

combining the closed-world submodel and the open-world submodel by utilizing the forecasted submodel weight.

18. The method of claim 15, generating the open-world submodel further comprising distributing probabilities associated with non-observed data based upon an analysis of spatial trends.

19. The method of claim 13, further comprising presenting relevant content based upon the prediction.

20. A system that facilitates generating a prediction, comprising:

means for generating a closed-world inference based upon observed data;

means for generating an open-world inference by evaluating a likelihood of observing previously non-observed data over a future period of time; and

means for combining the closed-world inference and the open-world inference based upon a weight to yield the prediction.