NEUROMORPHIC SYSTEM FOR OPERATING SPIKE-TIMING DEPENDENT PLASTICITY

摘要

The present invention relates to a neuromorphic system for synaptic learning in a neuromorphic array structure based on a spiking neural network (SNN). According to the present invention, by arranging a control block for implementing an STDP (Spike Timing Dependent Plasticity) operation including a post-synaptic neuron that generates a post-neuron spike on an output line of a synaptic array, synaptic learning can be stably implemented in an SNN neuromorphic array. In addition, by adding a lateral inhibition circuit, when a post-neuron spike is generated in an STDP control block connected to a certain output line, spikes are not generated in the STDP control blocks connected to the remaining output lines, thereby improving learning selectivity and, through this, improving the performance of the STDP algorithm.