Automatic model maintenance through local nets
First Claim
Patent Images
1. A model maintenance method comprising:
- collecting new data related to a system model, after the system model is formed;
determining that an accuracy of a prediction by the system model through consultation with the new data is below a predetermined threshold;
forming a compound model by supplementing the system model with a local net trained with the new data; and
storing the compound model on a storage device.
3 Assignments
0 Petitions
Accused Products
Abstract
A model maintenance method is provided. If accuracy of prediction by a current model through consultation with new data is determined to be below a predetermined threshold, a compound model is formed by supplementing the current model with a local net trained with the new data.
-
Citations
28 Claims
-
1. A model maintenance method comprising:
-
collecting new data related to a system model, after the system model is formed; determining that an accuracy of a prediction by the system model through consultation with the new data is below a predetermined threshold; forming a compound model by supplementing the system model with a local net trained with the new data; and storing the compound model on a storage device. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 28)
-
-
15. A computer system, comprising:
-
a processor; and a program storage device readable by the computer system, tangibly embodying a program of instructions executable by the processor to perform a model maintenance method, the method comprising; collecting new data related to a current model, after the current model is formed; determining that an accuracy of a prediction by the current model through consultation with the new data is below a predetermined threshold; forming a compound model by supplementing the current model with a local net trained with the new data; and storing the compound model.
-
-
16. A program storage device readable by a machine, tangibly embodying a program of instructions executable by the machine to perform a model maintenance method, the method comprising:
-
collecting new data related to a system model, after the system model is formed; determining that an accuracy of a prediction by a current model through consultation with the new data is below a predetermined threshold; forming a compound model by supplementing the current model with a local net trained with the new data; and storing the compound model.
-
-
17. A model maintenance method comprising:
-
determining that an accuracy of a current model is below a predetermined threshold; collecting data related to the current model, after the current model is formed, for adaptively updating the current model; forming a compound model by supplementing the current model with a local net trained with the collected data; and storing the compound model. - View Dependent Claims (18, 19, 20, 21)
-
-
22. A compound model of a system, comprising:
-
a current model; new data related to a current model, the new data collected after the current model is formed; at least one local net having an associated valid data space, wherein when the compound model is consulted with a data point of the new data, the local net is consulted with the data point and a result of consulting the local net is returned, if the data point is within an associated valid data space of the local net, and the current model is consulted with the data point and a result of consulting the current model is returned, if the data point is not within the associated valid data space of the local net. - View Dependent Claims (23, 24, 25)
-
-
26. A model maintenance method, comprising:
-
applying data to a system model to return a prediction; upon determining that the accuracy of the prediction of the system model is below a first threshold, training a first local net, the first local net having a first associated data space; supplementing the system model with the first local net to form a first compound model; storing the first compound model; and applying subsequent data to the first compound model, wherein; upon determining that a data point of the subsequent data is within the first associated data space of the first local net, consulting the first local net with the data point and returning a prediction; and upon determining that a data point of the subsequent data is not within the first associated data space of the first local net, consulting the system model with the data point and returning a prediction. - View Dependent Claims (27)
-
Specification