The further development of traditional von Neumann-architecture computers is limited by the breaking of Moore's law and the von Neumann bottleneck, which make them unsuitable for future high-performance artificial intelligence (AI) systems. Therefore, new computing paradigms are desperately needed. Inspired by the human brain, neuromorphic computing is proposed to realize AI while reducing power consumption. As one of the basic hardware units for neuromorphic computing, artificial synapses have recently aroused worldwide research interests. Among various electronic devices that mimic biological synapses, synaptic transistors show promising properties, such as the ability to perform signal transmission and learning simultaneously, allowing dynamic spatiotemporal information processing applications. In this article, we provide a review of recent advances in electrolyte- and ferroelectric-gated synaptic transistors. Their structures, materials, working mechanisms, advantages, and disadvantages will be presented. In addition, the challenges of developing advanced synaptic transistors are discussed.

With the need of the internet of things, big data, and artificial intelligence, creating new computing architecture is greatly desired for handling data-intensive tasks. Human brain can simultaneously process and store information, which would reduce the power consumption while improve the efficiency of computing. Therefore, the development of brain-like intelligent device and the construction of brain-like computation are important breakthroughs in the field of artificial intelligence. Memristor, as the fourth fundamental circuit element, is an ideal synaptic simulator due to its integration of storage and processing characteristics, and very similar activities and the working mechanism to synapses among neurons which are the most numerous components of the brains. In particular, memristive synaptic devices with optoelectronic responding capability have the benefits of storing and processing transmitted optical signals with wide bandwidth, ultrafast data operation speed, low power consumption, and low cross-talk, which is important for building efficient brain-like computing networks. Herein, we review recent progresses in optoelectronic memristor for neuromorphic computing, including the optoelectronic memristive materials, working principles, applications, as well as the current challenges and the future development of the optoelectronic memristor.

Phase-change material (PCM) is generating widespread interest as a new candidate for artificial synapses in bio-inspired computer systems. However, the amorphization process of PCM devices tends to be abrupt, unlike continuous synaptic depression. The relatively large power consumption and poor analog behavior of PCM devices greatly limit their applications. Here, we fabricate a GeTe/Sb₂Te₃ superlattice-like PCM device which allows a progressive RESET process. Our devices feature low-power consumption operation and potential high-density integration, which can effectively simulate biological synaptic characteristics. The programming energy can be further reduced by properly selecting the resistance range and operating method. The fabricated devices are implemented in both artificial neural networks (ANN) and convolutional neural network (CNN) simulations, demonstrating high accuracy in brain-like pattern recognition.

High-performance neuromorphic computing (i.e., brain-like computing) is envisioned to seriously demand optoelectronically integrated artificial neural networks (ANNs) in the future. Optoelectronic synaptic devices are critical building blocks for optoelectronically integrated ANNs. For the large-scale deployment of high-performance neuromorphic computing in the future, it would be advantageous to fabricate optoelectronic synaptic devices by using advanced silicon (Si) technologies. This calls for the development of Si-based optoelectronic synaptic devices. In this work we review the use of Si materials to make optoelectronic synaptic devices, which have either two-terminal or three-terminal structures. A series of important synaptic functionalities have been well mimicked by using these Si-based optoelectronic synaptic devices. We also present the outlook of using Si materials for optoelectronic synaptic devices.

Rapid developments in artificial intelligence trigger demands for perception and learning of external environments through visual perception systems. Neuromorphic devices and integrated system with photosensing and response functions can be constructed to mimic complex biological visual sensing behaviors. Here, recent progresses on optoelectronic neuromorphic memristors and optoelectronic neuromorphic transistors are briefly reviewed. A variety of visual synaptic functions stimulated on optoelectronic neuromorphic devices are discussed, including light-triggered short-term plasticities, long-term plasticities, and neural facilitation. These optoelectronic neuromorphic devices can also mimic human visual perception, information processing, and cognition. The optoelectronic neuromorphic devices that simulate biological visual perception functions will have potential application prospects in areas such as bionic neurological optoelectronic systems and intelligent robots.

Facing the computing demands of Internet of things (IoT) and artificial intelligence (AI), the cost induced by moving the data between the central processing unit (CPU) and memory is the key problem and a chip featured with flexible structural unit, ultra-low power consumption, and huge parallelism will be needed. In-memory computing, a non-von Neumann architecture fusing memory units and computing units, can eliminate the data transfer time and energy consumption while performing massive parallel computations. Prototype in-memory computing schemes modified from different memory technologies have shown orders of magnitude improvement in computing efficiency, making it be regarded as the ultimate computing paradigm. Here we review the state-of-the-art memory device technologies potential for in-memory computing, summarize their versatile applications in neural network, stochastic generation, and hybrid precision digital computing, with promising solutions for unprecedented computing tasks, and also discuss the challenges of stability and integration for general in-memory computing.

Memristive devices have attracted intensive attention in developing hardware neuromorphic computing systems with high energy efficiency due to their simple structure, low power consumption, and rich switching dynamics resembling biological synapses and neurons in the last decades. Fruitful demonstrations have been achieved in memristive synapses neurons and neural networks in the last few years. Versatile dynamics are involved in the data processing and storage in biological neurons and synapses, which ask for carefully tuning the switching dynamics of the memristive emulators. Note that switching dynamics of the memristive devices are closely related to switching mechanisms. Herein, from the perspective of switching dynamics modulations, the mainstream switching mechanisms including redox reaction with ion migration and electronic effect have been systemically reviewed. The approaches to tune the switching dynamics in the devices with different mechanisms have been described. Finally, some other mechanisms involved in neuromorphic computing are briefly introduced.

The multiple ferroelectric polarization tuned by external electric field could be used to simulate the biological synaptic weight. Ferroelectric synaptic devices have two advantages compared with other reported ones: One is that the intrinsic switching of ferroelectric domains without invoking of defect migration as in resistive oxides, contributes reliable performance in these ferroelectric synapses. Another tremendous advantage is the extremely low energy consumption because the ferroelectric polarization is manipulated by electric field which eliminates the Joule heating by current as in magnetic and phase change memories. Ferroelectric synapses have potential for the construction of low-energy and effective brain-like intelligent networks. Here we summarize recent pioneering work of ferroelectric synapses involving the structure of ferroelectric tunnel junctions (FTJs), ferroelectric diodes (FDs), and ferroelectric field effect transistors (FeFETs), respectively, and shed light on future work needed to accelerate their application for efficient neural network.

Recently, neuromorphic devices for artificial intelligence applications have attracted much attention. In this work, a three-terminal electrolyte-gated synaptic transistor based on NdNiO₃ epitaxial films, a typical correlated electron material, is presented. The voltage-controlled metal-insulator transition was achieved by inserting and extracting H⁺ ions in the NdNiO₃ channel through electrolyte gating. The non-volatile conductance change reached 10⁴ under a 2 V gate voltage. By manipulating the amount of inserted protons, the three-terminal NdNiO₃ artificial synapse imitated important synaptic functions, such as synaptic plasticity and spike-timing-dependent plasticity. These results show that the correlated material NdNiO₃ has great potential for applications in neuromorphic devices.

Emulation of synaptic function by ionic/electronic hybrid device is crucial for brain-like computing and neuromorphic systems. Electric-double-layer (EDL) transistors with proton conducting electrolytes as the gate dielectrics provide a prospective approach for such application. Here, artificial synapses based on indium-tungsten-oxide (IWO)-based EDL transistors are proposed, and some important synaptic functions (excitatory post-synaptic current, paired-pulse facilitation, filtering) are emulated. Two types of spike-timing-dependent plasticity (Hebbian STDP and anti-Hebbian STDP) learning rules and multistore memory (sensory memory, short-term memory, and long-term memory) are also mimicked. At last, classical conditioning is successfully demonstrated. Our results indicate that IWO-based neuromorphic transistors are interesting for neuromorphic applications.

The resistive random access memory (RRAM) has stimulated a variety of promising applications including programmable analog circuit, massive data storage, neuromorphic computing, etc. These new emerging applications have huge demands on high integration density and low power consumption. The cross-point configuration or passive array, which offers the smallest footprint of cell size and feasible capability of multi-layer stacking, has received broad attention from the research community. In such array, correct operation of reading and writing on a cell relies on effective elimination of the sneaking current coming from the neighboring cells. This target requires nonlinear I-V characteristics of the memory cell, which can be realized by either adding separate selector or developing implicit build-in nonlinear cells. The performance of a passive array largely depends on the cell nonlinearity, reliability, on/off ratio, line resistance, thermal coupling, etc. This article provides a comprehensive review on the progress achieved concerning 3D RRAM integration. First, the authors start with a brief overview of the associative problems in passive array and the category of 3D architectures. Next, the state of the arts on the development of various selector devices and self-selective cells are presented. Key parameters that influence the device nonlinearity and current density are outlined according to the corresponding working principles. Then, the reliability issues in 3D array are summarized in terms of uniformity, endurance, retention, and disturbance. Subsequently, scaling issue and thermal crosstalk in 3D memory array are thoroughly discussed, and applications of 3D RRAM beyond storage, such as neuromorphic computing and CMOL circuit are discussed later. Summary and outlooks are given in the final.