Analog neural network building blocks based on current mode subthreshold operation

A class of analog circuit building blocks for neural network implementation is presented which is based on the current mode subthreshold operation. This class of circuits includes current inverter, current-current multiplier, current amplifier and current comparator. A feedforward network with learning constructed by these building blocks was verified. These building blocks can be applied to artificial neural network (ANN) design in particular and to analog signal processing in general.<>