Skip to Main Content

Table 2 shows the models for developing SNNs—their architectures and learning type along with their accuracy rates on the MNIST data set. This comparison provides an insight into different SNN architectures and learning mechanisms to choose the right tool for the right purpose in future investigations.

Table 2:

Summary of Recent SNN Learning Models and Their Accuracy on Handwritten Digits Data Set (NIST).

ReferenceNetwork-TypeEncoding MethodStructure ConfigurationNeuron TypeLearning TypeLearning RuleTraining/Test SampleCA (%)
Mirsadeghi et al. (2021) SNN Temporal coding 784-500-10 Linear SRM Supervised STiDi-BP 60,000/10,000 97.4 
Fu & Dong (2021) Spiking CNN Rank order coding NA LIF Unsupervised Variable threshold + STD 60,000/10,000 99.27 
Qu et al. (2020) SNN Temporal coding 784-400-400-10 Nonleaky IF Unsupervised STDP 60,000/10,000 92 
Xu et al. (2020) Deep CovDenseSNN Rate coding 6C6@28×28-12C5-24C5-P LIF Unsupervised Hybrid spike-based learning. STDP 60,000/10,000 91.4 
Xu et al. (2020) Spiking CNN Rate coding 5C5-2P-64C5-2P-10FC IF Supervised Conversion rule 70,000 99.09 
Wang et al. (2020) Deep SNN Rate coding 64C3-MP2-64C3-2MP-128FC-10 LIF Supervised Weights-threshold balance conversion 60,000/10,000 99.43 
Zhou et al. (2019) Spiking CNN Temporal coding NA Nonleaky IF Supervised Temporal backpropagation 60,000/10,000 98.50 
Zheng and Mazumder (2018a) SNN Rate coding 784-300-100-10 LIF Supervised Online learning stochastic GD 60,000/10,000 97.8 
Kulkarni and Rajendran (2018) SNN NA 784-8112-10 LIF Supervised Normalized approximate descent 50,000/10,000 98.17 
Lee et al. (2018) Deep CovDenseSNN Rate coding 28×28-20C5-2P-50C5-2P-200FC-10FC LIF Semisupervised STDP-based pretraining + backpropagation 60,000/10,000 99.28 
Shrestha and Orchard (2018) Deep SNN Temporal coding 28×28-12c5-2a-64c5-2a-10o LIF Supervised Backpropagation 60,000/10,000 99.36 
Tavanaei et al. (2018) Spiking CNN Temporal coding 64C5-2P-1500FC-10FC LIF Both STDP rep. learning and BP-STDP 60,000/10,000 98.60 
Mostafa (2017) SNN Temporal coding 784-400-400-10 Nonleaky IF Supervised Temporal backpropagation 60,000/10,000 97.14 
Xu et al. (2017) Spiking ConvNet Rate coding 28×28-32c5-2s-64c5-2s-1024f-10o IF Supervised Conversion rule 20,000 99.17 
Stromatias et al. (2017) Spiking CNN Temporal coding NA LIF Supervised Stochastic GD 60,000/10,000 98.42 
ReferenceNetwork-TypeEncoding MethodStructure ConfigurationNeuron TypeLearning TypeLearning RuleTraining/Test SampleCA (%)
Mirsadeghi et al. (2021) SNN Temporal coding 784-500-10 Linear SRM Supervised STiDi-BP 60,000/10,000 97.4 
Fu & Dong (2021) Spiking CNN Rank order coding NA LIF Unsupervised Variable threshold + STD 60,000/10,000 99.27 
Qu et al. (2020) SNN Temporal coding 784-400-400-10 Nonleaky IF Unsupervised STDP 60,000/10,000 92 
Xu et al. (2020) Deep CovDenseSNN Rate coding 6C6@28×28-12C5-24C5-P LIF Unsupervised Hybrid spike-based learning. STDP 60,000/10,000 91.4 
Xu et al. (2020) Spiking CNN Rate coding 5C5-2P-64C5-2P-10FC IF Supervised Conversion rule 70,000 99.09 
Wang et al. (2020) Deep SNN Rate coding 64C3-MP2-64C3-2MP-128FC-10 LIF Supervised Weights-threshold balance conversion 60,000/10,000 99.43 
Zhou et al. (2019) Spiking CNN Temporal coding NA Nonleaky IF Supervised Temporal backpropagation 60,000/10,000 98.50 
Zheng and Mazumder (2018a) SNN Rate coding 784-300-100-10 LIF Supervised Online learning stochastic GD 60,000/10,000 97.8 
Kulkarni and Rajendran (2018) SNN NA 784-8112-10 LIF Supervised Normalized approximate descent 50,000/10,000 98.17 
Lee et al. (2018) Deep CovDenseSNN Rate coding 28×28-20C5-2P-50C5-2P-200FC-10FC LIF Semisupervised STDP-based pretraining + backpropagation 60,000/10,000 99.28 
Shrestha and Orchard (2018) Deep SNN Temporal coding 28×28-12c5-2a-64c5-2a-10o LIF Supervised Backpropagation 60,000/10,000 99.36 
Tavanaei et al. (2018) Spiking CNN Temporal coding 64C5-2P-1500FC-10FC LIF Both STDP rep. learning and BP-STDP 60,000/10,000 98.60 
Mostafa (2017) SNN Temporal coding 784-400-400-10 Nonleaky IF Supervised Temporal backpropagation 60,000/10,000 97.14 
Xu et al. (2017) Spiking ConvNet Rate coding 28×28-32c5-2s-64c5-2s-1024f-10o IF Supervised Conversion rule 20,000 99.17 
Stromatias et al. (2017) Spiking CNN Temporal coding NA LIF Supervised Stochastic GD 60,000/10,000 98.42 

Close Modal

or Create an Account

Close Modal
Close Modal