Spike-timing dependent plasticity in a transistor-selected resistive switching memory.

摘要: In a neural network, neuron computation is achieved through the summation of input signals fed by synaptic connections. The activity (weight) dictated synchronous firing neurons, inducing potentiation/depression connection. This learning function can be supported resistive switching memory (RRAM), which changes its resistance depending on amplitude, pulse width and bias polarity applied signal. work shows new synapse circuit comprising MOS transistor as selector RRAM variable resistance, displaying spike-timing dependent plasticity (STDP) similar to one originally experienced in biological networks. We demonstrate long-term potentiation depression simulations with an analytical model switching. Finally, experimental demonstration STDP scheme presented.