Pulse mode multilayer neural network based on floating point number representation

This paper proposes a new type of pulse mode multilayer neural network (MNN) that uses floating point number system for synapse weight value. Combined with pulse mode operation, the floating point operation is implemented without a multiplier. The proposed neuron, synapse multiplier and experimental MNN are implemented on field programmable gate arrays (FPGA) and various experiments are conducted to test the feasibility of the system. These experiments prove that the proposed MNN architecture can be used for applications which require high precision in their calculation