Quasi-Volatile MoS₂ Barristor Memory for 1T Compact Neuron by Correlative Charges Trapping and Schottky Barrier Modulation

Efficient Spiking Neural Networks with Biologically Similar Lithium-Ion Memristor Neurons

“…Spike neural networks (SNN), benefiting from the event-driven nature and the sparse asynchronous binary peaking coding, are considered as potential alternatives to ANNs for ultralow power consumption. − Neural models have been proposed for the SNN as McCulloch–Pitts (MP), , Hodgkin–Huxley (H–H), , integrate-and-fire (IF), − and leak-integrate-and-fire (LIF). − Among the models, the LIF model, a representative synaptic element, plays an important role in the construction of SNN systems because it can simultaneously deal with the spatial and temporal integration of input signals, continuously achieve the leakage and firing function, and even mimic the threshold dynamics of biological neurons. , Importantly, artificial synapses combined with peripheral circuits can build LIF models that replicate similar structures of biological neurons. , …”