Neural network learning system
AssigneeFujitsu Limited
InventorMasayuki Yokono
Life of this patent
Estimating the term…
Abstract
A neural network learning system is applied to extensive use in applications such as pattern and character recognizing operations, various controls, etc. The neural network learning system operates on, for example, a plurality of neural networks each having a different number of intermediate layer units to efficiently perform a learning process at a high speed with a reduced amount of hardware. A neural network system having a plurality of hierarchical neural networks each having an input layer, one or more intermediate layers and output layers is formed from a common input layer shared among two or more neural networks, or the common input layer and one or more intermediate layers and a learning controller for controlling a learning process performed by a plurality of neural networks.
Claims6
A neural network learning system having a plurality of hierarchical neural networks each having an input layer, at least one intermediate layer and an output layer, said learning system comprising:
- a common input layer unit commonly used by the hierarchical neural networks as the input layer;
- at least one intermediate layer unit, coupled to said common input layer unit, to provide the at least one intermediate layer of the hierarchical neural networks;
- output layer units coupled to said intermediate layer unit, each providing an output of the output layer of one of the hierarchical neural networks;
- a learning control unit, coupled to said intermediate and output layer units, for controlling a learning process performed by the hierarchical neural networks;
- a pattern storing unit, coupled to said common input layer unit, for storing at least one input pattern for said common input layer unit used during the learning process, and at least one teaching pattern corresponding to a desired output from a corresponding neural network in response to the at least one input pattern;
- and
- a plurality of error calculating units, each coupled to said pattern storing unit, a corresponding output layer unit and said learning control unit, for calculating an error between the output of each output layer unit and the teaching pattern corresponding thereto, and for outputting the error to said learning control unit.
Dependencies are inferred from the claim text.