Neurobiological systems use impressively few resources of energy, space, and time to solve complex sensory and sensorimotor tasks. An important reason for such efficiency is the clever use of nonlinear, adaptive, distributed, and hybrid analog-digital computational strategies. First, I will describe ongoing research on building high-performance, ultra-low-power bionic-ear processors inspired by the biological cochlea. Then, I describe why the optimal strategy for efficient computation is likely to be a hybrid mixture of analog and digital computation. I conclude by outlining research on building energy-efficient architectures that are inspired by pulsatile analog-digital representations in the brain's neurons.
Copyright. All Rights Reserved.