Memristive Synapses for Brain‐Inspired Computing

Abstract

be precisely regulated by the ionic flow, which is called synaptic plasticity. Generally, there are two types of synaptic plasticity: potentiation and depression. In the case of potentiation or depression, the synaptic weight increases or decreases upon repeated stimulation. Inspired by the human brain, novel non von Neumann computing architectures mainly composed of artificial synapses and artificial neurons have been proposed, which we can call "artificial neural networks," "brain-like computers," or "neuromorphic computers." [1,[3][4][5] However, the development of artificial brains that possess the efficiency of their biological counterparts in information processing remains a major challenge in computing. [6] One of the major challenges is the lack of suitable hardware to implement synapses, which are the main data processing elements in artificial brains. Generally, there are four essential requirements for an artificial synapse. [7,8] First, the analog weight should be nonvolatile in the absence of learning and weight updates should be bidirectional when the synapse is learning. Second, the synapse output is the product of the input signal and the synapse weight. Third, the weight updates vary with both the input signal and the stored weight value. Fourth, the synapse should be rather compact, and operate off a unidirectional information transfer with low power consumption.Mead and co-workers fabricated a four-terminal synaptic device based on a single floating gate silicon transistor in 1995. [8] It can simultaneously perform long term weight storage, compute the product of the input and floating gate value, and update the weight value according to a Hebbian or a backpropagation learning rule. In 1996, they developed a new floating-gate silicon transistor synapse with a threeterminal structure. [7] Hot-electron injection and electron tunneling make possible bidirectional weight updates. Given that these updates depend on both the stored weight value and the transistor terminal voltages, such synapse can implement a learning function. However, the integration density of transistor synapses is severely restricted to their three-or fourterminal structures. [6] The emergence of memristors endows artificial synapses with a new development opportunity. The memristor (short for memory resistor) is a two-terminal circuit element characterized by a relationship between the charge and the flux-linkage, which shows a distinctive "fingerprint" characterized by a pinched hysteresis loop confined to the first and the third quadrants of the Although the structure and function of the human brain are still far from being fully understood, brain-inspired computing architectures mainly consisting of artificial neurons and artificial synapses have been attracting more and more attentions due to their powerful computing capability and energy efficient operation. Synaptic plasticity is believed to be the origin of learning and memory. However, it is still a big challenge to realize artificial synapses with high reliability, go...

Memristive Synapses and Neurons for Bioinspired Computing

“…[57] Then, memristive devices have been extensively studied over the past decade due to their prominent advantages, such as simple structure, high operation speed, and low power consumption in applications of data storage, logic operation, and neuromorphic computation. [11] In addition, the resistive switching behavior can also be classified into volatile and nonvolatile types based on the state retention characteristics. The resistance of memristive device can be reversely changed between low resistive state (LRS) and high resistive state (HRS), resulting in a pinched hysteresis current-voltage (I-V) loop by applying voltage sweeping.…”

Photonic Synapses for Ultrahigh‐Speed Neuromorphic Computing

“…It is favorable for high‐density integration of synaptic devices. In the case of the sandwich structured device with a transparent electrode, the crossbar array architecture can be used to integrate a huge number of artificial synapses, in which each crosspoint represents a synaptic device . However, sneak paths, an inherent disadvantage of crossbar arrays, will result in an incorrect device conductance readout.…”

“…Furthermore, the size of a crossbar array, i.e., maximum numbers of rows and columns, is severely limited by the existence of sneak paths . Several techniques can be adopted to address the sneak path problem such as nonlinear or self‐rectifying resistance‐switching behavior of synaptic devices and integrating a transistor with each synapse …”

“…The deep neural network is trained according to a general algorithm by which the weight of each edge of the network can be optimized using an error‐correction function or maximum likelihood function . However, the operation of deep neural network based on modern computer systems is severely restricted due to the von Neumann architecture in which the main memory and central processing unit are physically separated . The data transfer process spends an enormous amount of energy and time.…”

“…Energy is consumed only when and where it is needed to process the information. Hence, brain‐inspired computing or neuromorphic computing based on a non‐von Neumann architecture is believed to be a promising way for the efficient implementation of artificial intelligence . A neuromorphic computing system is mainly composed of artificial synapses and neurons.…”