Knowledge Representation and Learning Mechanism Based on Networks of Spiking Neurons

Abstract

The function of acknowledging the fundamental data processing within the brain poses, aside from several experimental queries, demands theoretical issues on all levels, from molecules to behaviors. This chapter focuses on modeling techniques at the extent of neurons and few populations of neurons since we expect that this is often a suitable level to deal with principle queries of neuronal coding, the transmission of signals, or synaptic plasticity. In this chapter, succeeding topics have been shown: spiking neuron model and spike-timing-dependent plasticity (STDP) learning, rule representation based on spiking neural networks (SNNs), and waveform matching learning-SNN (WML-SNN) model. SNN constitutes a particular category of artificial neural networks, which the neuron model transmits with the aid of using a sequence of spikes. Networks composed of spiking neurons are up to a position to produce a massive quantity of statistics using a relatively small variety of spikes. On account of their functional comparability to biological neurons, spiking methods offer useful gear for the evolution of primary methods inside the brain and neural information processing, plasticity, and learning. STDP learning capacity has been seen in manual memristors, if the STDP is brought about on the memristor or the neuron is indistinct. This topic demonstrated the STDP attribute in the model for both asymmetric and symmetric memristors and utilized the asymmetric/symmetric memristors with STDP property and the streamlined neurons to play out the STDP learning capacity.