MNIST Recognition Using Unsupervised Biologically Learning

In this paper, spiking neural networks (SNN) inspired by the model of local cortical population as a biologicalneuro-computing resource for digit recognition was presented. SNN was equipped with spike-based unsupervisedweight optimization based on the dynamical behavior of the excitatory (AMPA) and inhibitory (GABA) synapsesusing Spike Timing Dependent Plasticity (STDP). There are two main reasons why this structure is state of the artcompared to previous works: learning process is compatible with many experimental observations on induction oflong-term potentiation and long-term depression, image to signal mapping created an informative signal of theimage based on sequences of prolate spheroidal wave functions (PSWFs). Cortical SNN compared toearlier related studies recognized MNIST digits more accurate and achieved 96.1% classificationaccuracy with unsupervised learning based on sparse spike activity.

Keywords:

Spiking Neural Network , STDP , MNIST Recognition , Biologically Learning