Learning system for a data processing apparatus
First Claim
1. A learning system for use in a data processing apparatus for learning an input pattern by obtaining an internal-state value necessary for realizing a desired data conversion, comprising:
- a pattern converting neural network having a multiplier, an accumulator and a threshold value processor for performing a pattern conversion defined by the internal-state value and for calculating an output pattern corresponding to the input pattern;
pattern presenting means, coupled to said pattern converting neural network, for inputting to the learning system an input pattern group of a subject to be learned for pattern conversion, said pattern presenting means dividing the input pattern group of the subject to be learned into a least two sets, selecting one of the divided sets, and presenting the input patterns belonging to all the sets presented up to the current point when an internal-state value to be converged is obtained in accordance with the input of said selected set;
error value calculating mans for calculating an error value representing a magnitude of an inconsistency between an output pattern outputted from said pattern presenting means in accordance with the presented patterns and a teacher pattern representing a pattern designated by said output; and
an internal-state value updating means for updating an internal-state value in accordance with an updated quantity of an internal-state value calculated based on the error value and obtaining an internal-state value having an error value within a specified allowance.
0 Assignments
0 Petitions
Accused Products
Abstract
A learning system is used in a data processing apparatus for learning an input pattern by obtaining an internal-state value necessary for realizing a desired data conversion by performing a pattern conversion defined by the internal-state value and calculating an output pattern corresponding to the input pattern. The learning system comprises a pattern presenting unit for presenting an input pattern group of the subject to be learned for pattern conversion, dividing the input pattern group of the subject to be learned into at least two sets, selecting one of the divided sets, presenting the input pattern group of the selected set to a pattern conversion unit and presenting an input pattern group belonging to all the sets presented up to the current point when the internal-state value to be converged is obtained in accordance with the presentation of the selected set, and an error value calculating unit for calculating an error value representing a magnitude of a non-consistency between an output pattern group outputted in accordance with the presentation and a teacher pattern group representing a pattern to be obtained by the output pattern group.
24 Citations
8 Claims
-
1. A learning system for use in a data processing apparatus for learning an input pattern by obtaining an internal-state value necessary for realizing a desired data conversion, comprising:
-
a pattern converting neural network having a multiplier, an accumulator and a threshold value processor for performing a pattern conversion defined by the internal-state value and for calculating an output pattern corresponding to the input pattern; pattern presenting means, coupled to said pattern converting neural network, for inputting to the learning system an input pattern group of a subject to be learned for pattern conversion, said pattern presenting means dividing the input pattern group of the subject to be learned into a least two sets, selecting one of the divided sets, and presenting the input patterns belonging to all the sets presented up to the current point when an internal-state value to be converged is obtained in accordance with the input of said selected set; error value calculating mans for calculating an error value representing a magnitude of an inconsistency between an output pattern outputted from said pattern presenting means in accordance with the presented patterns and a teacher pattern representing a pattern designated by said output; and an internal-state value updating means for updating an internal-state value in accordance with an updated quantity of an internal-state value calculated based on the error value and obtaining an internal-state value having an error value within a specified allowance. - View Dependent Claims (2, 3, 4, 5, 6)
-
-
7. A learning system for use in a data processing apparatus for learning an input pattern by obtaining an internal-state value necessary for realizing a desired data conversion, performing a pattern conversion defined by the internal-state value, and calculating an output pattern corresponding to the input pattern, said system comprising:
-
an adaptable data processing apparatus including; an internal-state value storing unit for performing a data processing function; and a pattern converting neural network including a multiplier, an accumulator and a threshold value processor, operatively connected to said internal-state value storing unit and operatively connected to receive an input pattern, for converting the input pattern to an output pattern in accordance with the data processing function of said internal-state value storing unit; learning pattern storing means for storing groups of learning patterns necessary to learn an internal-state value, including; input storing means for storing a plurality of input patterns for the learning process; and teacher pattern storing means for storing teacher patterns provided as output patterns corresponding to the input patterns; a pattern presentation unit, operatively connected between said learning pattern storing unit and said adaptable data processing apparatus, for reading a group of learning patterns from said learning pattern storing means, inputting the input patterns to said pattern converting neural network, and for distributing said teacher patterns among the input patterns, including; presentation pattern determining means for dividing the learning pattern groups into at least two sets and determining the pattern group to be presented; presentation-pattern transmitting means for performing a presentation process of the learning pattern groups determined by said presentation-pattern determining means; and presentation-amplified pattern-management means for managing the learning pattern group; an error value calculating unit, operatively connected between said pattern presentation unit and said adaptable data processing apparatus, for calculating an error represented by an amplitude of a non-coincidence between the output patterns and the teacher patterns. - View Dependent Claims (8)
-
Specification