Neuromorphic Engineering: From Biological to Spike‐Based Hardware Nervous Systems

Abstract

The human brain is a sophisticated, high-performance biocomputer that processes multiple complex tasks in parallel with high efficiency and remarkably low power consumption. Scientists have long been pursuing an artificial intelligence (AI) that can rival the human brain. Spiking neural networks based on neuromorphic computing platforms simulate the architecture and information processing of the intelligent brain, providing new insights for building AIs. The rapid development of materials engineering, device physics, chip integration, and neuroscience has led to exciting progress in neuromorphic computing with the goal of overcoming the von Neumann bottleneck. Herein, fundamental knowledge related to the structures and working principles of neurons and synapses of the biological nervous system is reviewed. An overview is then provided on the development of neuromorphic hardware systems, from artificial synapses and neurons to spike-based neuromorphic computing platforms. It is hoped that this review will shed new light on the evolution of brain-like computing.

Dual‐Mode Learning of Ambipolar Synaptic Phototransistor Based on 2D Perovskite/Organic Heterojunction for Flexible Color Recognizable Visual System

“…[ 1–5 ] The neuron functions simultaneously as a processor and a memory unit, and exhibits high parallelism, low power consumption, and fault tolerance. [ 6–8 ] However, the internal processer and memory of the traditional von Neumann computer are physically separated from each other, which takes a lot of time to transfer data, resulting in very low real‐tivme processing efficiency for visual information perception and complex speech recognition. [ 9–11 ] To solve this problem and to fulfill our increasing desire for data and information, various synaptic devices have been developed, such as memristors, [ 12 ] resistive switching memory, [ 13 ] and field effect synaptic transistor.…”

Bio‐Inspired Photoelectric Artificial Synapse based on Two‐Dimensional Ti3C2Tx MXenes Floating Gate

“…For the past decades, with the explosive growth in demand for big data processing, the conventional computing technology based on complementary silicon metal-oxide-semiconductor investigated. [5,[18][19][20] Compared with the two-terminal synaptic memristors, the three-terminal structured synaptic transistors are believed to benefit the control of synaptic weight due to the avoidance of setting training and testing inputs in the same terminal. [21][22][23] Typically, for synaptic transistors, the applied electrical input on the gate electrode (V GS ) or the optical input on the channel is regarded as the training signals.…”

Bidirectional Light-Driven Ion Transport through Porphyrin Metal–Organic Framework-Based van der Waals Heterostructures via pH-Induced Band Alignment Inversion

“…54 With enhanced stimulus intensity, both forward and reverse pH polarization generate growing output synaptic weights (Figure 6d), which is analogy to the potentiation behavior of synaptic strength and obeys Hebbian learning theory. 55 Switching between positive and negative potentiation states can be realized by controlling the pH polarization. This feature is different from previous artificial synapses that both potentiation and depression are found under electric or optical stimulus.…”