Iterative transfer of machine-trained network inputs from validation set to training set
AssigneePerceive Corporation
InventorsSteven L. Teig, Eric A. Sather
Life of this patent
Estimating the term…
Abstract
Some embodiments provide a method for training a machine-trained (MT) network. The method uses a first set of training inputs to train parameters of the MT network. The method uses a set of validation inputs to measure error for the MT network as trained by the first set of training inputs. The method adds at least a subset of the validation inputs to the first set of training inputs to create a second set of training inputs. The method uses the second set of training inputs to train the parameters of the MT network. The error measurement is used to modify the training with the second set of training inputs.
Claims20
A method for training a machine-trained (MT) network, the method comprising:
- using a first set of training inputs to train parameters of the MT network;
- using a first set of validation inputs to measure error for the MT network as trained by the first set of training inputs;
- moving a first subset of the first set of validation inputs to the first set of training inputs to create a second set of training inputs and a second set of validation inputs, wherein (i) the second set of training inputs comprises the first set of training inputs and the first subset of the first set of validation inputs and (ii) the second set of validation inputs comprises a second subset of the first set of validation inputs;
- using the second set of training inputs to train the parameters of the MT network, wherein the error measurement determined using the first set of validation inputs is used to modify the training with the second set of training inputs;
- and
- using the second set of validation inputs to measure error for the MT network as trained by the second set of training inputs, wherein a plurality of inputs are used in both the first and second sets of validation inputs to measure the error for the MT network as trained by the first and second sets of training inputs respectively.
A non-transitory machine-readable medium storing a program which when executed by at least one processing unit trains a machine-trained (MT) network, the program comprising sets of instructions for:
- using a first set of training inputs to train parameters of the MT network;
- using a first set of validation inputs to measure error for the MT network as trained by the first set of training inputs;
- moving a first subset of the first set of validation inputs to the first set of training inputs to create a second set of training inputs and a second set of validation inputs, wherein (i) the second set of training inputs comprises the first set of training inputs and the first subset of the first set of validation inputs and (ii) the second set of validation inputs comprises a second subset of the first set of validation inputs;
- using the second set of training inputs to train the parameters of the MT network, wherein the error measurement determined using the first set of validation inputs is used to modify the training with the second set of training inputs;
- and
- using the second set of validation inputs to measure error for the MT network as trained by the second set of training inputs, wherein a plurality of inputs are used in both the first and second sets of validation inputs to measure the error for the MT network as trained by the first and second sets of training inputs respectively.
Dependencies are inferred from the claim text.