I review the advancements of atomic scale nanoelectronics towards quantumneuromorphics. First, I summarize the key properties of elementary combinationsof few neurons, namely long-- and short--term plasticity, spike-timingdependent plasticity (associative plasticity), quantumness and stochasticeffects, and their potential computational employment. Next, I review severalatomic scale device technologies developed to control electron transport at theatomic level, including single atom implantation for atomic arrays and CMOSquantum dots, single atom memories, Ag$_2$S and Cu$_2$S atomic switches,hafnium based RRAMs, organic material based transistors, Ge$_2$Sb$_2$Te$_5$synapses. Each material/method proved successful in achieving some of theproperties observed in real neurons. I compare the different methods towardsthe creation of a new generation of naturally inspired and biophysicallymeaningful artificial neurons, in order to replace the rigid CMOS basedneuromorphic hardware. The most challenging aspect to address appears to obtainboth the stochastic/quantum behavior and the associative plasticity, which arecurrently observed only below and above 20 nm length scale respectively, byemploying the same material.
展开▼
