Intralayer-Connected Spiking Neural Network with Hybrid Training Using Backpropagation and Probabilistic Spike-Timing Dependent Plasticity

Table 2

Comparison of different models on the MNIST dataset.

Models	Structure		Average accuracy (%)	Max accuracy (%)
Wu et al. [11]	Net1		—	99.42
Jin et al. [36]	Net1	400	99.42	99.49
Zhang and Peng [6]	Net1	5	99.50	99.53
Wang et al. [13]	Net1	5	99.48	99.52
Zhang and Peng [12]	Net1	400	99.57	99.62
Amiri et al. [37]	—	—	—	97.8
ISNN	Net1	5	99.53	99.55
ISNN-HY	Net1	5	99.55	99.58

The bold values in the first column are the names of the proposed models. The bold values in the last two columns are the best average and max accuracies.