Abstract A method of accelerating the training of an artificial neural network uses a computer configured as an artificial neural network with a network input and a network output and having a plurality of interconnected units arranged in layers including an input layer and an output layer. Each unit has a multiplicity of unit inputs and a set of variables for operating upon a unit inputs to provide a unit output in the range between binary 1 and binary 0. A plurality of training examples is serially provided to the network input and the network output is observed. The computer is programmed with a back propagation algorithm for changing each set of variables in response to feedback representing differences between the network output for each example and the desired output. The examples are iterated while the output of a unit is observed. The feedback to a unit is adjusted so that a larger feedback is obtained when the output of the unit is near binary 1 or binary 0 than when the output is midrange between binary 1 or binary 0.
A method of accelerating the training of an artificial neural network comprising the steps of: providing a computer configured as an artificial neural network with a network input and a network output and having a plurality of interconnected units arranged in layers including an input layer and an output layer, each unit having a multiplicity of unit inputs and a set of variables for operating upon said unit inputs to provide a unit output in the range between binary 1 and binary 0; serially providing a plurality of training examples to said network input and observing the network output; programming said computer with a back propagation algorithm for changing each set of variables in response to feedback representing differences between the network output for each example and a desired output; iterating said examples; observing the output of a unit; adjusting the feedback to a unit so that a larger feedback is obtained when the output of the unit is near binary 1 or binary 0 than when the output is midrange between binary 1 or binary 0.
Claim 2
Independent Claim
177 words
A method of accelerating the training of an artificial neural network comprising the steps of: providing a computer configured as an artificial neural network with a network input and a network output and having a plurality of interconnected units arranged in layers including an input layer and an output layer, each unit having a multiplicity of unit inputs and a set of variables for operating upon said unit inputs to provide a unit output in the range between binary 1 and binary 0; serially providing a plurality of training examples to said network input and observing the network output; programming said computer with a back propagation algorithm for changing each set of variables in response to feedback representing differences between the network output for each example and a desired output; iterating said examples; observing the output of a unit; adjusting the feedback to a unit so that a larger feedback is obtained when the output of the unit is near binary 1 or binary 0 than when the output is midrange between binary 1 or binary 0.