Emergent Device Boosts Neuromorphic Computing

Los Alamos National Laboratory scientists have developed artificial synaptic devices that mimic brain functions, such as learning and memory. The interfacial properties between metal and dielectric layers determine the devices' resistance states, offering advantages over conventional computing devices like analogue-type memory, uniform switching, low power consumption, and high scalability. Electric pulses can control the devices, and researchers simulated an artificial neural network using these devices to recognize Modified National Institute of Standards and Technology (MNIST) digital images with high accuracy.

Read the full article here.