Span: Spike Pattern Association Neuron for Learning Spatio-Temporal Spike Patterns

Abstract

Spiking Neural Networks (SNN) were shown to be suitable tools for the processing of spatio-temporal information. However, due to their inherent complexity, the formulation of efficient supervised learning algorithms for SNN is difficult and remains an important problem in the research area. This article presents SPAN - a spiking neuron that is able to learn associations of arbitrary spike trains in a supervised fashion allowing the processing of spatio-temporal information encoded in the precise timing of spikes. The idea of the proposed algorithm is to transform spike trains during the learning phase into analog signals so that common mathematical operations can be performed on them. Using this conversion, it is possible to apply the well-known Widrow-Hoff rule directly to the transformed spike trains in order to adjust the synaptic weights and to achieve a desired input/output spike behavior of the neuron. In the presented experimental analysis, the proposed learning algorithm is evaluated regarding its learning capabilities, its memory capacity, its robustness to noisy stimuli and its classification performance. Differences and similarities of SPAN regarding two related algorithms, ReSuMe and Chronotron, are discussed.

Evaluating SPAN Incremental Learning for Handwritten Digit Recognition

“…In this paper we have demonstrated the application of SNN trained with SPAN [10][11][12] on learning and classifying images of handwritten digits. One crucial factor in using SNN for real-world computer application is properly encoding the information into spike patterns.…”

“…More details can be found in previous publications [10][11][12]. SPAN rule is a supervised learning method to associate input spike pattern to a target spike train by adjusting the weights of the input synapses according to the following formula:…”

“…According to previous studies such as in [6] and recent one [14], temporal coding whereby information is encoded into precise time of the spikes, plays a significant role in the neural code of the brain especially in the visual system. SPAN [12,11], is a learning algorithm for spiking neural networks which is based on encoding input information as precise time of spikes (Temporal coding). This is opposite to rate coding where information is coded in the mean firing rate of the neurons.…”

“…This is opposite to rate coding where information is coded in the mean firing rate of the neurons. The algorithm was evaluated mainly on two tasks: precise time spike sequence generation, and spike pattern classification [12]. In spike sequence generation task, a spiking neuron is trained to generate any random spike train in response to a recognised pattern of input spike sequences (spike trains).…”

“…

Abstract: In a previous work [12,11], the authors proposed SPAN: a learning algorithm based on temporal coding for Spiking Neural Network (SNN). The algorithm trains a neuron to associate target spike patterns to input spatio-temporal spike patterns.

…”