Emergent brain-like complexity from nanowire atomic switch networks: Towards neuromorphic synthetic intelligence

Abstract

__The atomic switch is a novel nanotechnology that mimics the chemical synapse between neurons in response to electrical stimuli. When connected together in a self- organized manner, similar to a neuronal network, atomic switch networks exhibit emergent brain-like complexity properties, including nonlinear stochastic dynamics and memorization, making them a unique experimental system for emulating intelligence. Here we present a computational model developed to simulate atomic switch networks to explore the scope of emergent brain-like features. Our modelling results demonstrate the capacity for neuromorphic atomic switch networks to emulate long-term memory and generate scale-invariant fluctuations in signal transmission, in direct analogy to the brain.