Implementation of hardware-based neural network using memristors with abrupt SET and gradual RESET characteristics

Recent advance of artificial intelligence comes from the rapid growth of artificial neural network technology. Memristors play a crucial role in the hardware implementation of artificial neural networks, thanks to the multilevel conductance of memristors by switching behaviors. Here, we propose a synaptic device with a Pr0.7Ca0.3MnO3 (PCMO) switching layer which shows the abrupt SET and gradual RESET switching characteristics. Improved linearity of the synaptic transmission caused by switching characteristics can enhance the classification performance of neuromorphic systems.