A Primer on Neuromorphic Hardware: Underlying Principles and Approaches—From Sensors to Processors
摘要
Brain-like energy efficiency and brain-like cognitive ability has intrigued researchers to mimic the fundamental computing principles of the brain on hardware computing platforms. Such neuromorphic systems present a paradigm shift in hardware computing technology and leverage both CMOS and beyond-CMOS emerging technologies to create large-scale neuromorphic systems. This chapter presents spike-based computations and spike-based communications as key fundamental principles underlying neuromorphic systems and discusses several approaches spanning sensors and processors that implements these fundamental principles into scalable, energy efficient and intelligent neuromorphic hardware solutions.