Probabilistic Evolving Spiking Neural Network Optimization Using Dynamic Quantum-inspired Particle Swarm Optimization

Abstract

This paper proposes a novel Probabilistic Evolving Spiking Neural Network (PESNN) based on Kasabov’s Probabilistic Neuron Model. The features, connections and parameters are optimized using Dynamic Quantum- inspired Particle Swarm Optimization (DQiPSO). The features and connections are modeled as a quantum bit vector while the parameter values are presented as real numbers. An improved search strategy is also being introduced to probe the most relevant features and eliminate irrelevant ones. The proposed method is evaluated using a synthetic dataset for classification problems. The results show that the proposed method is promising, with better accuracy and capability to identify the most significant features while obtaining the best combination of PESNN’s connections and parameters.