作者: Nan Du , Mahdi Kiani , Christian G. Mayr , Tiangui You , Danilo Bürger
关键词: Electronic engineering 、 Neuromorphic engineering 、 Waveform 、 Time constant 、 Voltage spike 、 Spike-timing-dependent plasticity 、 Memristor 、 Weight change 、 Neuroscience 、 Voltage 、 Computer science
摘要: Memristive devices are popular among neuromorphic engineers for their ability to emulate forms of spike-driven synaptic plasticity by applying specific voltage and current waveforms at two terminals. In this paper, we investigate spike-timing dependent (STDP) with a single pairing one presynaptic spike postsynaptic in BiFeO3 memristive device. most materials the learning window is primarily function material characteristics not applied waveform. contrast, show that analog resistive switching developed artificial synapses allows adjust time constant STDP from 25ms 125μs via duration spikes. Also, as induced weight change may degrade, remanence resistance several hours after switching, thus emulating processes expected biological synapses. As power consumption major constraint circuits, methods reduce consumed energy per setting pulse only 4.5 pJ