A Supervised Learning Algorithm for Learning Precise Timing of Multiple Spikes in Multilayer Spiking Neural Networks

Abstract

There is a biological evidence to prove information is coded through precise timing of spikes in the brain. However, training a population of spiking neurons in a multilayer network to fire at multiple precise times remains a challenging task. Delay learning and the effect of a delay on weight learning in a spiking neural network (SNN) have not been investigated thoroughly. This paper proposes a novel biologically plausible supervised learning algorithm for learning precisely timed multiple spikes in a multilayer SNNs. Based on the spike-timing-dependent plasticity learning rule, the proposed learning method trains an SNN through the synergy between weight and delay learning. The weights of the hidden and output neurons are adjusted in parallel. The proposed learning method captures the contribution of synaptic delays to the learning of synaptic weights. Interaction between different layers of the network is realized through biofeedback signals sent by the output neurons. The trained SNN is used for the classification of spatiotemporal input patterns. The proposed learning method also trains the spiking network not to fire spikes at undesired times which contribute to misclassification. Experimental evaluation on benchmark data sets from the UCI machine learning repository shows that the proposed method has comparable results with classical rate-based methods such as deep belief network and the autoencoder models. Moreover, the proposed method can achieve higher classification accuracies than single layer and a similar multilayer SNN.

Unsupervised Learning of a Hierarchical Spiking Neural Network for Optical Flow Estimation: From Events to Global Motion Perception

“…Several lines of research can be distinguished regarding the use of supervised learning in SNNs, with the most promising based on the well-known error backpropagation algorithm [36]. Firstly, numerous adaptations to the discontinuous dynamics of SNNs have recently been proposed for learning temporally precise spike patterns [37], [38], [39], [40]. Alternatively, due to the popularity of this method in ANNs, SNNs commonly rely on transferring optimization results from their non-spiking counterparts [41], [42], [43].…”

An Approach of One-vs-Rest Filter Bank Common Spatial Pattern and Spiking Neural Networks for Multiple Motor Imagery Decoding

“…SNN is proposed based on biological principles, which could simulate the connection and communication between neurons to the great extent. The spike neuron model is the mathematical abstraction of the real neuron, which shows good performance and strong robustness in some pattern recognition problems [13], [17], [34], [35]. Due to such advantage, SNN is used in this study to classify the optimal feature set of MI.…”

Optimization of Output Spike Train Encoding for a Spiking Neuron Based on its Spatio–Temporal Input Pattern

“…ReSuMe (Remote Supervised Method) [7] is a biologically plausible learning algorithm that works based on Spike Timing Dependent Plasticity (STDP) and anti-STDP to train a neuron firing desired output spikes at non-adapted times. QuickProp [8], RProp [9], Chronotron [10], SPAN [11], EMPD [12], BPSL [13], EDL [14], and the supervised method proposed in [15] are other examples of learning algorithms for training spiking neurons to fire at non-adapted desired output times. The times of desired output spikes are usually set randomly, and the random target spikes might not be an appropriate choice for a classification task, which can result in a reduction in learning efficiency.…”