Pulse pattern training of spiking neural networks using improved genetic algorithm

Tuning the parameters of an integrate and fire neuron via a genetic algorithm for solving pattern recognition problems

“…Pattern recognition will be carried out with the procedure described in [9][10][11][12][13][14][15][16]. The classification is made by counting the Table 2 Spike quantity for all the inputs and the parameter combination in Section 2.3.2 for T ¼100 ms. number of spikes generated, by each of the input patterns, when processed by the neuron.…”

“…was optimized, one can see that the initial error is normally smaller, this is because the fixed values for the parameters were chosen from already known values which worked well for pattern recognition [9][10][11][12][13][14][15][16].…”

“…Evolutionary Computation methods have been also applied during training. In [12], the authors described how a genetic algorithm can be used during the learning process of a spiking neural network; however, the model was not applied to perform a pattern recognition problem. In [13], the authors show how a spiking neural network can be trained by a quantum PSO and then applied to a string pattern recognition problem.…”