RRAM-based Analog-Weight Spiking Neural Network Accelerator with in-situ Learning for IoT Applications

With the rapid growth of data in the Internet of Things (IoT), more biologically realistic and computationally efficient spiking neural networks (SNNs) have been applied to energy-efficient intelligent edge devices. Accelerators with in-situ learning function can adapt to the dynamically changing environment by adjusting network parameters. However, current neuromorphic digital computing systems u...