Multi-layer perceptron circuit device

Claim 1

A multi-layer perceptron circuit device with a learning function, comprising: at least one tiling synapse block including: a plurality of synapses for performing weight calculation on synapse input signals to obtain output signals; synapse input signal lines for transmitting the synapse input signals to the synapses; synapse output signal lines for transmitting the synapse output signal from the synapses; learning means for updating values of weights to be utilized in weight calculation by a multiplication of synapse error input signals and the synapse input signals; synapse error signal input lines for transmitting the synapse error input signals to the synapses; synapse error signal output lines for transmitting the synapse error output signals from the synapses; and means for calculating the synapse error output signals by a multiplication of the synapse error input signals and the weights; at least one tiling input neuron block containing a plurality of neurons, each neuron of the at least one tiling input neuron block being connected with one of the synapse input signal lines by its signal output, inputting one of a plurality of perceptron input signals by its signal input, and supplying one of the synapse input signals to one of the synapse input signal lines, each neuron of the tiling input neuron blocks being connected to one of the synapse error signal output lines by its error signal input, receiving one of the synapse error output signals from one of the synapse error signal output lines, and outputting one of a plurality of perceptron error signals by its error signal output, each of the neurons of the tiling input neuron blocks further comprises: differential function converter means for converting one of the perceptron input signals into a converted signal using a nonlinear function; means for obtaining one of the perceptron error signals from one of the synapse error output signals transmitted through one of the synapse error signal output lines and the converted signal obtained by the differential function converter means; function converter means for converting one of the perceptron input signals inputted through said each of the neurons into one of the synapse input signals using a nonlinear function; and gain control circuit means for controlling a gain of the function converter means and the differential function converter means wherein said gain control circuit means further comprises means for continuously changing the gain of the function converter means and the differential function converter means in a process of learning by the learning means from a value different from that required for a desired result to a value required for the desired result; and at least one tiling output neuron block containing a plurality of neurons, each neuron of the tiling output neuron blocks being connected with one of the synapse output signal lines by its signal input, and receiving one of the synapse output signals from one of the synapse output signal lines, and outputting a perceptron output signal by its signal output, each neuron of the tiling output neuron blocks being connected to one of the synapse error signal input lines by its error signal output, inputting an externally supplied teacher signal by its error signal input, and supplying one of the synapse error input signals to one of the synapse error signal input lines.