Research on the Effect of Firing Threshold on Spike-Based Context Dependent Learning Network

Context-dependent learning allows the network model to learn and adjust autonomously according to the environment. SNN allows the network to run at high speed and low power consumption. The selection of the threshold value allows the network to achieve the highest correctness and the fastest response time. Through the observation of experimental correctness, response time, SI place, SI item, SI context, and Binariness, it is concluded that when the threshold value of this model is 5, the experimental correctness is highest and the fastest reaction time can be achieved by choosing different thresholds for different situations. This experimental result lays the foundation and guarantee for the future establishment of other models to directly find the discharge threshold with the highest experimental correctness or to seek for the fastest experimental response time.