Ferroelectric photosensor network: an advanced hardware solution to real-time machine vision

Binary Neural Network Based on a Programmable Graphene/Si Schottky Diode for In-Sensor Processing Image Sensors

Recent advancements in in-sensor computing technology have demonstrated significant advantages in time latency and energy efficiency in visual information processing through device-level integration of photosensing and neuromorphic computing. However, current implementations face challenges due to their single-layer architecture, creating an urgent demand for the development of devices that integrate front-end in-sensor processing with back-end computing layers. Here, we report a programmable graphene/Si Schottky diode (PGSSD) featuring gate-voltage-programmed photoresponsivity and rectification direction. The programmability of the photoresponsivity enables the application of reconfigurable convolution kernels to implement in-sensor convolution of optical images. Simultaneously, the programmable rectification direction permits analog-domain execution of quasi-binary multiply-accumulate (MAC) operations. Based on these capabilities, we constructed a complete binary neural network (BNN) using the PGSSDs and demonstrated its application for image recognition. The BNN combines front-end convolution processing and back-end computing layers, achieving an inference accuracy of 98.35% on the MNIST database.

Polarization-independent narrowband photodetection with plasmon-induced thermoelectric effect in a hexagonal array of Au nanoholes

Photodetectors are crucial for modern technologies such as optical communications, imaging, autonomous vehicles, and machine vision. However, conventional semiconductor-based photodetectors require additional filtering systems due to their broad spectral response, leading to increased costs and complexity. Here, we present a narrow spectral response photodetector using hexagonally arranged plasmonic Au nanohole structures, eliminating the need for optical filters. The device achieves a full-width at half maximum (FWHM) bandwidth of ∼40 nm with a response peak at 760 nm and a linear photocurrent responsivity of 0.95 μA/W. The photothermoelectric effect, induced by the nonradiative decay of plasmonic resonance, converts optical radiation into an electric potential on the Au surface. The hexagonal nanohole design generates polarization-independent photocurrents and allows spectral tuning beyond the cutoff region of silicon photodetectors. This versatile approach enables customizable response characteristics across a broad wavelength range through geometric design, enhancing its potential for diverse applications.

Reconfigurable Phototransistor Based on the Interface Trap State of the WS2/Ta2NiSe5 Heterostructure for Logic-in-Memory Operation

In recent years, machine vision technology has experienced rapid growth, in which optoelectronic sensing plays a crucial role. With the rising demand for versatile integration of photoelectric sensors, sensors that combine sensing, storage, and computing capabilities are emerging as the next-generation paradigm for high-density, low-power visual processing. Here, with a simplified floating-gate structure based on the interface trap state of the WS2/Ta2NiSe5 heterostructure, we demonstrate a self-driven logic-in-memory optoelectronic transistor with nonvolatile reconfigurable photoresponsivity by controlling the direction and magnitude of the built-in electric field, which can be used for image preprocessing, at the same time capable of achieving multifunctions such as neuromorphic computation, multicolor sensing, positive and negative light response switching, and polarization detection, highlighting the great potential for multifunctional reconfigurable vision devices.

In-Sensor Polarization Convolution Based on Ferroelectric-Reconfigurable Polarization-Sensitive Photodiodes

In-sensor computing can enhance the imaging system performance by putting part of the computations into the sensor. While substantial advancements have been made in latency, spectral range, and functionalities, the strategy for in-sensor light polarization computing has remained unexplored. Here, it is shown that ferroelectric-reconfigurable polarization-sensitive photodiodes (FPPDs) based on BiFeO3 nanowire arrays can perform in-sensor computations on polarization information. This innovation leverages the anisotropic photoresponse from the 1D structure of nanowires and the non-volatile reconfigurability of ferroelectrics. The devices show programmable anisotropic ratios as high as 5219, surpassing most state-of-the-art polarization-sensitive photodetectors and commercial polarization image sensors. Employing tunable photoresponse as kernel, FPPDs can perform convolutions to directly extract feature maps containing polarization information, which raises the recognition accuracy on road-scene objects under adverse weather up to 89.6%. The research highlights the potential of FPPDs as a highly efficient vision sensor and extends the boundaries of advanced intelligent imaging systems.

In-memory ferroelectric differentiator

Differential calculus is the cornerstone of many disciplines, spanning the breadth of modern mathematics, physics, computer science, and engineering. Its applications are fundamental to theoretical progress and practical solutions. However, the current state of digital differential technology often requires complex implementations, which struggle to meet the extensive demands of the ubiquitous edge computing in the intelligence age. To face these challenges, we propose an in-memory differential computation that capitalizes on the dynamic behavior of ferroelectric domain reversal to efficiently extract information differences. This strategy produces differential information directly within the memory itself, which considerably reduces the volume of data transmission and operational energy consumption. We successfully illustrate the effectiveness of this technique in a variety of tasks, including derivative function solving, the moving object extraction and image discrepancy identification, using an in-memory differentiator constructed with a crossbar array of 1600-unit ferroelectric polymer capacitors. Our research offers an efficient hardware analogue differential computing, which is crucial for accelerating mathematical processing and real-time visual feedback systems.

Bipolar-Response Perovskite Photodetector Controlled by the Ferroelectric Depolarization Field for Secure Optical Communication

Traditional optical communication systems are constrained by fixed photoresponse values and light intensity, significantly impairing the potential for data transmission and protection. Here, a single-band bipolar-response perovskite self-powered photodetector is demonstrated on the (PEA)2PbI4/BaTiO3/Si heterojunction. By employing BaTiO3 as the intrinsic layer, the device demonstrates a low dark current on the order of 10-12 A at a 5 V bias. When BaTiO3 functions as the ferroelectric layer, the variation in the depolarization field not only achieves multilevel modulation of the photoresponse magnitude but also reverses the sign. Leveraging the bipolar characteristics of the device, a secure optical communication system has been developed, featuring a dual-channel optical signal reception specifically designed for information transmission. One channel is designated for receiving encrypted information, while the other channel receives key information. The device only needs to identify the positive and negative values of the input signals with arbitrary light intensity without distinguishing the strength of the signal values. The accurate retrieval of transmitted information is contingent upon the application of an encryption algorithm, thereby enhancing the security of the communication system. This work provides novel perspectives for the realization of more secure and reliable encrypted optical communication systems.

High-Precision Attention Mechanism for Machine Vision Enabled by an Artificial Optoelectronic Memristor Synapse

The rapid advancement of artificial intelligence has facilitated the broad application of machine vision systems in diverse industries. However, these systems are often confronted with computational challenges stemming from an overwhelming amount of data. Here, we have developed a novel optoelectronic memristor synapse constructed from an ITO/Nb:SrTiO3 heterostructure, which synergistically integrates light signal detection with information processing and memory functions. Notably, we have achieved precise decoupling of the interactions between light power and wavelength at the hardware level, significantly enhancing the accuracy and efficiency of image processing. Furthermore, by incorporating an attention mechanism analogous to that of human vision, we have enabled the device to weight key information and filter out irrelevant data. Experimental results demonstrate that this attention mechanism can increase the accuracy of facial recognition by 13% while reducing the data load by 35-65%. This work is expected to advance the development of optoelectronic synapses in machine vision.

Ferroelectric Optoelectronic Memory Based on p-GaN/ZnGa2O4/BaTiO3/n-ITO Heterojunction with Integrated Sensing and Logic Operations

Ferroelectric optoelectronic memories, capable of integrating sensing, computing, and storage functionalities, hold significant potential in the fields of artificial intelligence and the Internet of Things. In this study, a nonvolatile p-GaN/ZnGa2O4/BaTiO3/n-ITO ferroelectric optoelectronic memory is demonstrated. By combining with wide-bandgap semiconductors ZnGa2O4 and GaN, known for their excellent optoelectronic properties, the device exhibits superior self-powered ultraviolet photodetection performance. At 0 V bias, the device achieves a peak responsivity of 7 mA/W with a fast response speed (rise time: 6 ms and fall time: 12 ms). Furthermore, by adjusting the polarization direction of the BaTiO3 thin film, the optoelectronic performance of the device can be modulated to achieve memory functionality. The photocurrent of the device in the up-polarized state remains stable for over 5 months, indicating excellent long-term storage characteristics. Based on these properties, a 5 × 5 ferroelectric optoelectronic memory array capable of imaging, storing, and reading out has been demonstrated. Additionally, the device can function as "AND" and "OR" logic gates depending on its initial polarization state and input signals. The results provide an avenue for the application of ferroelectric optoelectronic memory in integrated sensing, memory, and computing systems.

Critical band-to-band-tunnelling based optoelectronic memory

Neuromorphic vision hardware, embedded with multiple functions, has recently emerged as a potent platform for machine vision. To realize memory in sensor functions, reconfigurable and non-volatile manipulation of photocarriers is highly desirable. However, previous technologies bear mechanism challenges, such as the ambiguous optoelectronic memory mechanism and high potential barrier, resulting in a limited response speed and a high operating voltage. Here, for the first time, we propose a critical band-to-band tunnelling (BTBT) based device that combines sensing, integration and memory functions. The nearly infinitesimal barrier facilitates the tunnelling process, resulting in a broadband application range (940 nm). Furthermore, the observation of dual negative differential resistance (NDR) points confirms that the critical BTBT of photocarriers contributes to the sub-microsecond photomemory speed. Since the photomemory speed, with no motion blur, is important for motion detection, the critical BTBT memory is expected to enable moving target tracking and recognition, underscoring its superiority in intelligent perception.

Ion Intercalation-Enabled Reconfigurable Photodetector with Low Programming Voltage and Broadband Response Based on Van Der Waals Tin Disulfide

Artificial neural networks with integrated sensing and computing capabilities, leveraging reconfigurable optoelectronics, can effectively emulate biological neural networks, thereby enabling rapid and efficient information processing. However, realizing reconfigurable photoresponsivity is often blocked by the requirement for high programming voltages and the limits of the detection spectrum range. This greatly restricts the progress of energy-efficient and precise neuromorphic vision sensing. Herein, a reconfigurable photodetector with low programming voltage and broadband response is presented via in situ intercalation of Cu+ ions into the van der Waals (vdW) gaps of thermoelectric 2D material SnS2. Interestingly, the vdW gaps provide an ionic transport channel with lower energy barriers compared to oxide-based memristors, resulting in a low programming voltage (0.5 V). Furthermore, reversible conversion of photo-detection is achieved from photovoltaic to photo-thermoelectric (PTE) mode via voltage-controlled ion distribution, which modulates the phonon scattering rate in the neighboring SnS2 layer. As a result, the response spectrum switches from visible (532 nm) to long-wave infrared (10 µm) with an on/off ratio as high as 104. Thus, dual-mode conversion and broadband detection functionality in reconfigurable imaging are realized, suggesting a potential pathway for the development of highly energy-efficient reconfigurable optoelectronics with a spectrum far beyond human vision.

Ferroelectricity and Nonlinear Optical Responses in Two-Dimensional Distorted MX2Y (M = Cu, Ag, Au; X = Chalcogens; Y = Halogen) Monolayers

Two-dimensional (2D) materials with spontaneous polarization can exhibit large second-order nonlinear optical (NLO) effects. Here, we present a series of stable distorted MX2Y monolayers by using first-principles calculations and lattice vibration analysis. The structural distortion leads to a lower polar symmetry, giving rise to intrinsic ferroelectricity with a Curie point up to room temperature. We show that the bulk photovoltaic effect (BPVE) in polar MX2Y monolayers has an enhanced shift current, an order of magnitude larger than that in the undistorted nonpolar counterparts. Meanwhile, the second harmonic generation (SHG) susceptibility reaches up to the order of 106 pm2/V, superior to that of 2D conventional materials such as MoS2, h-BN, and GeS. Our study advances the research in 2D ferroelectric materials and would stimulate more efforts in developing optoelectronic devices based on NLO effects.

Optoelectronic array of photodiodes integrated with RRAMs for energy-efficient in-sensor computing

The rapid development of internet of things (IoT) urgently needs edge miniaturized computing devices with high efficiency and low-power consumption. In-sensor computing has emerged as a promising technology to enable in-situ data processing within the sensor array. Here, we report an optoelectronic array for in-sensor computing by integrating photodiodes (PDs) with resistive random-access memories (RRAMs). The PD-RRAM unit cell exhibits reconfigurable optoelectronic output and photo-responsivity by programming RRAMs into different resistance states. Furthermore, a 3 × 3 PD-RRAM array is fabricated to demonstrate optical image recognition, achieving a universal architecture with ultralow latency and low power consumption. This study highlights the great potential of the PD-RRAM optoelectronic array as an energy-efficient in-sensor computing primitive for future IoT applications.

In situ training of an in-sensor artificial neural network based on ferroelectric photosensors

In-sensor computing has emerged as an ultrafast and low-power technique for next-generation machine vision. However, in situ training of in-sensor computing systems remains challenging due to the demands for both high-performance devices and efficient programming schemes. Here, we experimentally demonstrate the in situ training of an in-sensor artificial neural network (ANN) based on ferroelectric photosensors (FE-PSs). Our FE-PS exhibits self-powered, fast (<30 μs), and multilevel (>4 bits) photoresponses, as well as long retention (50 days), high endurance (109), high write speed (100 ns), and small cycle-to-cycle and device-to-device variations (~0.66% and ~2.72%, respectively), all of which are desirable for the in situ training. Additionally, a bi-directional closed-loop programming scheme is developed, achieving a precise and efficient weight update for the FE-PS. Using this programming scheme, an in-sensor ANN based on the FE-PSs is trained in situ to recognize traffic signs for commanding a prototype autonomous vehicle. Moreover, this in-sensor ANN operates 50 times faster than a von Neumann machine vision system. This study paves the way for the development of in-sensor computing systems with in situ training capability, which may find applications in new data-streaming machine vision tasks.

Interface Charge Engineering in Ferroelectric Neuristors for a Complete Machine Vision System

The rapid advancement of artificial intelligence has driven the demand for hardware solutions of neuromorphic pathways to effectively mimic biological functions of the human visual system. However, current machine vision systems (MVSs) fail to fully replicate retinal functions and lack the ability to update weights through all-optical pulses. Here, by employing rational interface charge engineering via varying the charge trapping layer thickness of PMMA, we determine that the ferroelectric polarization of our ferroelectric neuristors can be flexibly manipulated through light or electrical pulses. This capability enables dynamic modulation of the device's optoelectronic characteristics, facilitating a complete MVS. As front-end sensors, devices with the thickest PMMA (∼32 nm) demonstrate autonomous light adaptation while those with the thinnest PMMA (∼2 nm) exhibit bidirectional photoresponse characteristics akin to those of bipolar cells. Furthermore, as components of a back-end processor, the conductances of these devices with a moderate thickness (∼12 nm) can be updated linearly through all-optical pulses. Our MVS, constructed with these neuristors, achieved an impressive recognition accuracy of 93% in handwritten digit recognition tasks under extreme lighting conditions. This work offers an effective strategy for the development of energy-efficient and highly integrated intelligent MVSs.

Object Motion Detection Enabled by Reconfigurable Neuromorphic Vision Sensor under Ferroelectric Modulation

Increasing the demand for object motion detection (OMD) requires shifts of reducing redundancy, heightened power efficiency, and precise programming capabilities to ensure consistency and accuracy. Drawing inspiration from object motion-sensitive ganglion cells, we propose an OMD vision sensor with a simple device structure of a WSe2 homojunction modulated by a ferroelectric copolymer. Under optical mode and intermediate ferroelectric modulation, the vision sensor can generate progressive and bidirectional photocurrents with discrete multistates under zero power consumption. This design enables reconfigurable devices to emulate long-term potentiation and depression for synaptic weights updating, which exhibit 82 states (more than 6 bits) with a uniform step of 6 pA. Such OMD devices also demonstrate nonvolatility, reversibility, symmetry, and ultrahigh linearity, achieving a fitted R2 of 0.999 and nonlinearity values of 0.01/-0.01. Thus, a vision sensor could implement motion detection by sensing only dynamic information based on the brightness difference between frames, while eliminating redundant data from static scenes. Additionally, the neural network utilizing a linear result can recognize the essential moving information with a high recognition accuracy of 96.8%. We also present the scalable potential via a uniform 3 × 3 neuromorphic vision sensor array. Our work offers a platform to achieve motion detection based on controllable and energy-efficient ferroelectric programmability.

Photoconductivity and Photovoltaic Effect Strengthened via Microstructural Cotuning in Ferroelectrics: Intuitively Assessed by Macroscopic Transparency

Photoferroelectrics that involve strong light-matter coupling are regarded as promising candidates for realizing bulk photovoltaic and photoelectric effects via light absorption. Nonetheless, understanding the photoresponse mechanism or modulation of performance from a microscopic point of view is scarcely explored through quantification of macroscopic properties. Herein, we design a model material, Gd3+-doped (K0.5Na0.5)NbO3 ferroelectric-transparent ceramics, and present an advantageous strategy to enhance the optoelectronic coupling through joint modulations of lattice distortion and oxygen vacancies, along with inner defects and ferroelectric domains. Significantly, their microcosmic manipulation can be intuitively and facilely evaluated by the optical transparency of each ceramic. An approximately 104 fold increase in conductivity under ultraviolet irradiation was produced. Under the cocoupling between external physical fields, the synergy of photoelectric stimulation increased the photoconductivity of the ceramics by 13.89 times. Additionally, a significant increase (4.5-fold) in the current output from the photovoltaic effect was achieved via ferroelectric domains of moderate size, whose size could be easily assessed by optical transmittance. In situ microscopic observations confirmed that the configuration of oxygen vacancy-dependent ferroelectric domains contributes to the enhanced optoelectronic response. This research provides a distinct way to develop inexpensive optocoupler devices and meet the requirements for multifunctional integration in single photoferroelectrics.

Producing Freestanding Single-Crystal BaTiO3 Films through Full-Solution Deposition

Strontium aluminate, with suitable lattice parameters and environmentally friendly water solubility, has been strongly sought for use as a sacrificial layer in the preparation of freestanding perovskite oxide thin films in recent years. However, due to this material's inherent water solubility, the methods used for the preparation of epitaxial films have mainly been limited to high-vacuum techniques, which greatly limits these films' development. In this study, we prepared freestanding single-crystal perovskite oxide thin films on strontium aluminate using a simple, easy-to-develop, and low-cost chemical full-solution deposition technique. We demonstrate that a reasonable choice of solvent molecules can effectively reduce the damage to the strontium aluminate layer, allowing successful epitaxy of perovskite oxide thin films, such as 2-methoxyethanol and acetic acid. Molecular dynamics simulations further demonstrated that this is because of their stronger adsorption capacity on the strontium aluminate surface, which enables them to form an effective protective layer to inhibit the hydration reaction of strontium aluminate. Moreover, the freestanding film can still maintain stable ferroelectricity after release from the substrate, which provides an idea for the development of single-crystal perovskite oxide films and creates an opportunity for their development in the field of flexible electronic devices.

Ferroelectric materials for neuroinspired computing applications

In recent years, the emergence of numerous applications of artificial intelligence (AI) has sparked a new technological revolution. These applications include facial recognition, autonomous driving, intelligent robotics, and image restoration. However, the data processing and storage procedures in the conventional von Neumann architecture are discrete, which leads to the "memory wall" problem. As a result, such architecture is incompatible with AI requirements for efficient and sustainable processing. Exploring new computing architectures and material bases is therefore imperative. Inspired by neurobiological systems, in-memory and in-sensor computing techniques provide a new means of overcoming the limitations inherent in the von Neumann architecture. The basis of neural morphological computation is a crossbar array of high-density, high-efficiency non-volatile memory devices. Among the numerous candidate memory devices, ferroelectric memory devices with non-volatile polarization states, low power consumption and strong endurance are expected to be ideal candidates for neuromorphic computing. Further research on the complementary metal-oxide-semiconductor (CMOS) compatibility for these devices is underway and has yielded favorable results. Herein, we first introduce the development of ferroelectric materials as well as their mechanisms of polarization reversal and detail the applications of ferroelectric synaptic devices in artificial neural networks. Subsequently, we introduce the latest developments in ferroelectrics-based in-memory and in-sensor computing. Finally, we review recent works on hafnium-based ferroelectric memory devices with CMOS process compatibility and give a perspective for future developments.

Ultralow Power In-Sensor Neuronal Computing with Oscillatory Retinal Neurons for Frequency-Multiplexed, Parallel Machine Vision

In-sensor and near-sensor computing architectures enable multiply accumulate operations to be carried out directly at the point of sensing. In-sensor architectures offer dramatic power and speed improvements over traditional von Neumann architectures by eliminating multiple analog-to-digital conversions, data storage, and data movement operations. Current in-sensor processing approaches rely on tunable sensors or additional weighting elements to perform linear functions such as multiply accumulate operations as the sensor acquires data. This work implements in-sensor computing with an oscillatory retinal neuron device that converts incident optical signals into voltage oscillations. A computing scheme is introduced based on the frequency shift of coupled oscillators that enables parallel, frequency multiplexed, nonlinear operations on the inputs. An experimentally implemented 3 × 3 focal plane array of coupled neurons shows that functions approximating edge detection, thresholding, and segmentation occur in parallel. An example of inference on handwritten digits from the MNIST database is also experimentally demonstrated with a 3 × 3 array of coupled neurons feeding into a single hidden layer neural network, approximating a liquid-state machine. Finally, the equivalent energy consumption to carry out image processing operations, including peripherals such as the Fourier transform circuits, is projected to be <20 fJ/OP, possibly reaching as low as 15 aJ/OP.

Multimodal Artificial Synapses for Neuromorphic Application

The rapid development of neuromorphic computing has led to widespread investigation of artificial synapses. These synapses can perform parallel in-memory computing functions while transmitting signals, enabling low-energy and fast artificial intelligence. Robots are the most ideal endpoint for the application of artificial intelligence. In the human nervous system, there are different types of synapses for sensory input, allowing for signal preprocessing at the receiving end. Therefore, the development of anthropomorphic intelligent robots requires not only an artificial intelligence system as the brain but also the combination of multimodal artificial synapses for multisensory sensing, including visual, tactile, olfactory, auditory, and taste. This article reviews the working mechanisms of artificial synapses with different stimulation and response modalities, and presents their use in various neuromorphic tasks. We aim to provide researchers in this frontier field with a comprehensive understanding of multimodal artificial synapses.

Dual-Adaptive Heterojunction Synaptic Transistors for Efficient Machine Vision in Harsh Lighting Conditions

Photoadaptive synaptic devices enable in-sensor processing of complex illumination scenes, while second-order adaptive synaptic plasticity improves learning efficiency by modifying the learning rate in a given environment. The integration of above adaptations in one phototransistor device will provide opportunities for developing high-efficient machine vision system. Here, a dually adaptable organic heterojunction transistor as a working unit in the system, which facilitates precise contrast enhancement and improves convergence rate under harsh lighting conditions, is reported. The photoadaptive threshold sliding originates from the bidirectional photoconductivity caused by the light intensity-dependent photogating effect. Metaplasticity is successfully implemented owing to the combination of ambipolar behavior and charge trapping effect. By utilizing the transistor array in a machine vision system, the details and edges can be highlighted in the 0.4% low-contrast images, and a high recognition accuracy of 93.8% with a significantly promoted convergence rate by about 5 times are also achieved. These results open a strategy to fully implement metaplasticity in optoelectronic devices and suggest their vision processing applications in complex lighting scenes.

Si/CuO Heterojunction-Based Photomemristor for Reconfigurable, Non-Volatile, and Self-Powered In-Sensor Computing

In-sensor computing has attracted considerable interest as a solution for overcoming the energy efficiency and response time limitations of the traditional von Neumann architecture. Recently, emerging memristors based on transition-metal oxides (TMOs) have attracted attention as promising candidates for in-memory computing owing to their tunable conductance, high speed, and low operational energy. However, the poor photoresponse of TMOs presents challenges for integrating sensing and processing units into a single device. This integration is crucial for eliminating the need for a sensor/processor interface and achieving energy-efficient in-sensor computing systems. In this study, a Si/CuO heterojunction-based photomemristor is proposed that combines the reversible resistive switching behavior of CuO with the appropriate optical absorption bandgap of the Si substrate. The proposed photomemristor demonstrates a simultaneous reconfigurable, non-volatile, and self-powered photoresponse, producing a microampere-level photocurrent at zero bias. The controlled migration of oxygen vacancies in CuO result in distinct energy-band bending at the interface, enabling multiple levels of photoresponsivity. Additionally, the device exhibits high stability and ultrafast response speed to the built-in electric field. Furthermore, the prototype photomemristor can be trained to emulate the attention-driven nature of the human visual system, indicating the tremendous potential of TMO-based photomemristors as hardware foundations for in-sensor computing.

A Reconfigurable Bipolar Image Sensor for High-Efficiency Dynamic Vision Recognition

Dynamic vision perception and processing (DVPP) is in high demand by booming edge artificial intelligence. However, existing imaging systems suffer from low efficiency or low compatibility with advanced machine vision techniques. Here, we propose a reconfigurable bipolar image sensor (RBIS) for in-sensor DVPP based on a two-dimensional WSe2/GeSe heterostructure device. Owing to the gate-tunable and reversible built-in electric field, its photoresponse shows bipolarity as being positive or negative. High-efficiency DVPP incorporating front-end RBIS and back-end CNN is then demonstrated. It shows a high recognition accuracy of over 94.9% on the derived DVS128 data set and requires much fewer neural network parameters than that without RBIS. Moreover, we demonstrate an optimized device with a vertically stacked structure and a stable nonvolatile bipolarity, which enables more efficient DVPP hardware. Our work demonstrates the potential of fabricating DVPP devices with a simple structure, high efficiency, and outputs compatible with advanced algorithms.

Inorganic Halide Perovskite Nanowires/Conjugated Polymer Heterojunction-Based Optoelectronic Synaptic Transistors for Dynamic Machine Vision

Inspired by the retina, artificial optoelectronic synapses have groundbreaking potential for machine vision. The field-effect transistor is a crucial platform for optoelectronic synapses that is highly sensitive to external stimuli and can modulate conductivity. On the basis of the decent optical absorption, perovskite materials have been widely employed for constructing optoelectronic synaptic transistors. However, the reported optoelectronic synaptic transistors focus on the static processing of independent stimuli at different moments, while the natural visual information consists of temporal signals. Here, we report CsPbBrI2 nanowire-based optoelectronic synaptic transistors to study the dynamic responses of artificial synaptic transistors to time-varying visual information for the first time. Moreover, on the basis of the dynamic synaptic behavior, a hardware system with an accuracy of 85% is built to the trajectory of moving objects. This work offers a new way to develop artificial optoelectronic synapses for the construction of dynamic machine vision systems.

Silicon-Compatible Ferroelectric Tunnel Junctions with a SiO2/Hf0.5Zr0.5O2 Composite Barrier as Low-Voltage and Ultra-High-Speed Memristors

The big data era requires ultrafast, low-power, and silicon-compatible materials and devices for information storage and processing. Here, ferroelectric tunnel junctions (FTJs) based on SiO2/Hf0.5Zr0.5O2 composite barrier and both conducting electrodes are designed and fabricated on Si substrates. The FTJ achieves the fastest write speed of 500 ps under 5 V (2 orders of magnitude faster than reported silicon-compatible FTJs) or 10 ns speed at a low voltage of 1.5 V (the lowest voltage among FTJs at similar speeds), low write current density of 1.3 × 104 A cm-2, 8 discrete states, good retention > 105 s at 85 °C, and endurance > 107. In addition, it provides a large read current (88 A cm-2) at 0.1 V, 2 orders of magnitude larger than reported FTJs. Interestingly, in FTJ-based synapses, gradually tunable conductance states (128 states) with high linearity (<1) are obtained by 10 ns pulses of <1.2 V, and a high accuracy of 91.8% in recognizing fashion product images is achieved by online neural network simulations. These results highlight that silicon-compatible HfO2-based FTJs are promising for high-performance nonvolatile memories and electrical synapses.

Nonvolatile and reconfigurable two-terminal electro-optic duplex memristor based on III-nitride semiconductors

With the fast development of artificial intelligence (AI), Internet of things (IOT), etc, there is an urgent need for the technology that can efficiently recognize, store and process a staggering amount of information. The AlScN material has unique advantages including immense remnant polarization, superior temperature stability and good lattice-match to other III-nitrides, making it easy to integrate with the existing advanced III-nitrides material and device technologies. However, due to the large band-gap, strong coercive field, and low photo-generated carrier generation and separation efficiency, it is difficult for AlScN itself to accumulate enough photo-generated carriers at the surface/interface to induce polarization inversion, limiting its application in in-memory sensing and computing. In this work, an electro-optic duplex memristor on a GaN/AlScN hetero-structure based Schottky diode has been realized. This two-terminal memristor shows good electrical and opto-electrical nonvolatility and reconfigurability. For both electrical and opto-electrical modes, the current on/off ratio can reach the magnitude of 104, and the resistance states can be effectively reset, written and long-termly stored. Based on this device, the "IMP" truth table and the logic "False" can be successfully reproduced, indicating the huge potential of the device in the field of in-memory sensing and computing.

High performance artificial visual perception and recognition with a plasmon-enhanced 2D material neural network

The development of neuromorphic visual systems has recently gained momentum due to their potential in areas such as autonomous vehicles and robotics. However, current machine visual systems based on silicon technology usually contain photosensor arrays, format conversion, memory and processing modules. As a result, the redundant data shuttling between each unit, resulting in large latency and high-power consumption, seriously limits the performance of neuromorphic vision chips. Here, we demonstrate an artificial neural network (ANN) architecture based on an integrated 2D MoS2/Ag nanograting phototransistor array, which can simultaneously sense, pre-process and recognize optical images without latency. The pre-processing function of the device under photoelectric synergy ensures considerable improvement of efficiency and accuracy of subsequent image recognition. The comprehensive performance of the proof-of-concept device demonstrates great potential for machine vision applications in terms of large dynamic range (180 dB), high speed (500 ns) and low energy consumption per spike (2.4 × 10-17 J).

Technology and Integration Roadmap for Optoelectronic Memristor

Optoelectronic memristors (OMs) have emerged as a promising optoelectronic Neuromorphic computing paradigm, opening up new opportunities for neurosynaptic devices and optoelectronic systems. These OMs possess a range of desirable features including minimal crosstalk, high bandwidth, low power consumption, zero latency, and the ability to replicate crucial neurological functions such as vision and optical memory. By incorporating large-scale parallel synaptic structures, OMs are anticipated to greatly enhance high-performance and low-power in-memory computing, effectively overcoming the limitations of the von Neumann bottleneck. However, progress in this field necessitates a comprehensive understanding of suitable structures and techniques for integrating low-dimensional materials into optoelectronic integrated circuit platforms. This review aims to offer a comprehensive overview of the fundamental performance, mechanisms, design of structures, applications, and integration roadmap of optoelectronic synaptic memristors. By establishing connections between materials, multilayer optoelectronic memristor units, and monolithic optoelectronic integrated circuits, this review seeks to provide insights into emerging technologies and future prospects that are expected to drive innovation and widespread adoption in the near future.

PbS/CsPbBr3 Heterojunction for Broadband Neuromorphic Vision Sensing

Neuromorphic light sensors with analogue-domain image processing capability hold promise for overcoming the energy efficiency limitations and latency of von Neumann architecture-based vision chips. Recently, metal halide perovskites, with strong light-matter interaction, long carrier diffusion length, and exceptional photoelectric conversion efficiencies, exhibit reconfigurable photoresponsivity due to their intrinsic ion migration effect, which is expected to advance the development of visual sensors. However, suffering from a large bandgap, it is challenging to achieve highly tunable responsivity simultaneously with a wide-spectrum response in perovskites, which will significantly enhance the image recognition accuracy through the machine learning algorithm. Herein, we demonstrate a broadband neuromorphic visual sensor from visible (Vis) to near-infrared (NIR) by coupling all-inorganic metal halide perovskites (CsPbBr3) with narrow-bandgap lead sulfide (PbS). The PbS/CsPbBr3 heterostructure is composed of high-quality single crystals of PbS and CsPbBr3. Interestingly, the ion migration of CsPbBr3 with the implementation of an electric field induces the energy band dynamic bending at the interface of the PbS/CsPbBr3 heterojunction, leading to reversible, multilevel, and linearly tunable photoresponsivity. Furthermore, the reconfigurable and broadband photoresponse in the PbS/CsPbBr3 heterojunction allows convolutional neuronal network processing for pattern recognition and edge enhancements from the Vis to the NIR waveband, suggesting the great potential of the PbS/CsPbBr3 heterostructure in artificial intelligent vision sensing.

Hierarchies in Visual Pathway: Functions and Inspired Artificial Vision

The development of artificial intelligence has posed a challenge to machine vision based on conventional complementary metal-oxide semiconductor (CMOS) circuits owing to its high latency and inefficient power consumption originating from the data shuffling between memory and computation units. Gaining more insights into the function of every part of the visual pathway for visual perception can bring the capabilities of machine vision in terms of robustness and generality. Hardware acceleration of more energy-efficient and biorealistic artificial vision highly necessitates neuromorphic devices and circuits that are able to mimic the function of each part of the visual pathway. In this paper, we review the structure and function of the entire class of visual neurons from the retina to the primate visual cortex within reach (Chapter 2) are reviewed. Based on the extraction of biological principles, the recent hardware-implemented visual neurons located in different parts of the visual pathway are discussed in detail in Chapters 3 and 4. Furthermore, valuable applications of inspired artificial vision in different scenarios (Chapter 5) are provided. The functional description of the visual pathway and its inspired neuromorphic devices/circuits are expected to provide valuable insights for the design of next-generation artificial visual perception systems.

Toward grouped-reservoir computing: organic neuromorphic vertical transistor with distributed reservoir states for efficient recognition and prediction

Reservoir computing has attracted considerable attention due to its low training cost. However, existing neuromorphic hardware, focusing mainly on shallow-reservoir computing, faces challenges in providing adequate spatial and temporal scales characteristic for effective computing. Here, we report an ultra-short channel organic neuromorphic vertical transistor with distributed reservoir states. The carrier dynamics used to map signals are enriched by coupled multivariate physics mechanisms, while the vertical architecture employed greatly increases the feedback intensity of the device. Consequently, the device as a reservoir, effectively mapping sequential signals into distributed reservoir state space with 1152 reservoir states, and the range ratio of temporal and spatial characteristics can simultaneously reach 2640 and 650, respectively. The grouped-reservoir computing based on the device can simultaneously adapt to different spatiotemporal task, achieving recognition accuracy over 94% and prediction correlation over 95%. This work proposes a new strategy for developing high-performance reservoir computing networks.

Integrated In-Memory Sensor and Computing of Artificial Vision Based on Full-vdW Optoelectronic Ferroelectric Field-Effect Transistor

The development and application of artificial intelligence have led to the exploitation of low-power and compact intelligent information-processing systems integrated with sensing, memory, and neuromorphic computing functions. The 2D van der Waals (vdW) materials with abundant reservoirs for arbitrary stacking based on functions and enabling continued device downscaling offer an attractive alternative for continuously promoting artificial intelligence. In this study, full 2D SnS2 /h-BN/CuInP2 S6 (CIPS)-based ferroelectric field-effect transistors (Fe-FETs) and utilized light-induced ferroelectric polarization reversal to achieve excellent memory properties and multi-functional sensing-memory-computing vision simulations are designed. The device exhibits a high on/off current ratio of over 105 , long retention time (>104  s), stable cyclic endurance (>350 cycles), and 128 multilevel current states (7-bit). In addition, fundamental synaptic plasticity characteristics are emulated including paired-pulse facilitation (PPF), short-term plasticity (STP), long-term plasticity (LTP), long-term potentiation, and long-term depression. A ferroelectric optoelectronic reservoir computing system for the Modified National Institute of Standards and Technology (MNIST) handwritten digital recognition achieved a high accuracy of 93.62%. Furthermore, retina-like light adaptation and Pavlovian conditioning are successfully mimicked. These results provide a strategy for developing a multilevel memory and novel neuromorphic vision systems with integrated sensing-memory-processing.

Reconfigurable Neuromorphic Computing: Materials, Devices, and Integration

Neuromorphic computing has been attracting ever-increasing attention due to superior energy efficiency, with great promise to promote the next wave of artificial general intelligence in the post-Moore era. Current approaches are, however, broadly designed for stationary and unitary assignments, thus encountering reluctant interconnections, power consumption, and data-intensive computing in that domain. Reconfigurable neuromorphic computing, an on-demand paradigm inspired by the inherent programmability of brain, can maximally reallocate finite resources to perform the proliferation of reproducibly brain-inspired functions, highlighting a disruptive framework for bridging the gap between different primitives. Although relevant research has flourished in diverse materials and devices with novel mechanisms and architectures, a precise overview remains blank and urgently desirable. Herein, the recent strides along this pursuit are systematically reviewed from material, device, and integration perspectives. At the material and device level, one comprehensively conclude the dominant mechanisms for reconfigurability, categorized into ion migration, carrier migration, phase transition, spintronics, and photonics. Integration-level developments for reconfigurable neuromorphic computing are also exhibited. Finally, a perspective on the future challenges for reconfigurable neuromorphic computing is discussed, definitely expanding its horizon for scientific communities.

Ferroelectric-defined reconfigurable homojunctions for in-memory sensing and computing

Recently, the increasing demand for data-centric applications is driving the elimination of image sensing, memory and computing unit interface, thus promising for latency- and energy-strict applications. Although dedicated electronic hardware has inspired the development of in-memory computing and in-sensor computing, folding the entire signal chain into one device remains challenging. Here an in-memory sensing and computing architecture is demonstrated using ferroelectric-defined reconfigurable two-dimensional photodiode arrays. High-level cognitive computing is realized based on the multiplications of light power and photoresponsivity through the photocurrent generation process and Kirchhoff's law. The weight is stored and programmed locally by the ferroelectric domains, enabling 51 (>5 bit) distinguishable weight states with linear, symmetric and reversible manipulation characteristics. Image recognition can be performed without any external memory and computing units. The three-in-one paradigm, integrating high-level computing, weight memorization and high-performance sensing, paves the way for a computing architecture with low energy consumption, low latency and reduced hardware overhead.

Reconfigurable, non-volatile neuromorphic photovoltaics

The neural network image sensor-which mimics neurobiological functions of the human retina-has recently been demonstrated to simultaneously sense and process optical images. However, highly tunable responsivity concurrent with non-volatile storage of image data in the neural network would allow a transformative leap in compactness and function of these artificial neural networks. Here, we demonstrate a reconfigurable and non-volatile neuromorphic device based on two-dimensional semiconducting metal sulfides that is concurrently a photovoltaic detector. The device is based on a metal-semiconductor-metal (MSM) two-terminal structure with pulse-tunable sulfur vacancies at the M-S junctions. By modulating sulfur vacancy concentrations, the polarities of short-circuit photocurrent can be changed with multiple stable magnitudes. The bias-induced motion of sulfur vacancies leads to highly reconfigurable responsivities by dynamically modulating the Schottky barriers. A convolutional neuromorphic network is finally designed for image processing and object detection using the same device. The results demonstrated that neuromorphic photodetectors can be the key components of visual perception hardware.

Voltage-Mode Ferroelectric Synapse for Neuromorphic Computing

Ferroelectric materials with a modulable polarization extent hold promise for exploring voltage-driven neuromorphic hardware, in which direct current flow can be minimized. Utilizing a single active layer of an insulating ferroelectric polymer, we developed a voltage-mode ferroelectric synapse that can continuously and reversibly update its states. The device states are straightforwardly manifested in the form of variable output voltage, enabling large-scale direct cascading of multiple ferroelectric synapses to build a deep physical neural network. Such a neural network based on potential superposition rather than current flow is analogous to the biological counterpart driven by action potentials in the brain. A high accuracy of over 97% for the simulation of handwritten digit recognition is achieved using the voltage-mode neural network. The controlled ferroelectric polarization, revealed by piezoresponse force microscopy, turns out to be responsible for the synaptic weight updates in the ferroelectric synapses. The present work demonstrates an alternative strategy for the design and construction of emerging artificial neural networks.

Broadband sensory networks with locally stored responsivities for neuromorphic machine vision

As the most promising candidates for the implementation of in-sensor computing, retinomorphic vision sensors can constitute built-in neural networks and directly implement multiply-and-accumulation operations using responsivities as the weights. However, existing retinomorphic vision sensors mainly use a sustained gate bias to maintain the responsivity due to its volatile nature. Here, we propose an ion-induced localized-field strategy to develop retinomorphic vision sensors with nonvolatile tunable responsivity in both positive and negative regimes and construct a broadband and reconfigurable sensory network with locally stored weights to implement in-sensor convolutional processing in spectral range of 400 to 1800 nanometers. In addition to in-sensor computing, this retinomorphic device can implement in-memory computing benefiting from the nonvolatile tunable conductance, and a complete neuromorphic visual system involving front-end in-sensor computing and back-end in-memory computing architectures has been constructed, executing supervised and unsupervised learning tasks as demonstrations. This work paves the way for the development of high-speed and low-power neuromorphic machine vision for time-critical and data-intensive applications.

Emerging Iontronic Neural Devices for Neuromorphic Sensory Computing

Living organisms have a very mysterious and powerful sensory computing system based on ion activity. Interestingly, studies on iontronic devices in the past few years have proposed a promising platform for simulating the sensing and computing functions of living organisms, because: 1) iontronic devices can generate, store, and transmit a variety of signals by adjusting the concentration and spatiotemporal distribution of ions, which analogs to how the brain performs intelligent functions by alternating ion flux and polarization; 2) through ionic-electronic coupling, iontronic devices can bridge the biosystem with electronics and offer profound implications for soft electronics; 3) with the diversity of ions, iontronic devices can be designed to recognize specific ions or molecules by customizing the charge selectivity, and the ionic conductivity and capacitance can be adjusted to respond to external stimuli for a variety of sensing schemes, which can be more difficult for electron-based devices. This review provides a comprehensive overview of emerging neuromorphic sensory computing by iontronic devices, highlighting representative concepts of both low-level and high-level sensory computing and introducing important material and device breakthroughs. Moreover, iontronic devices as a means of neuromorphic sensing and computing are discussed regarding the pending challenges and future directions.

Silicon based Bi0.9La0.1FeO3 ferroelectric tunnel junction memristor for convolutional neural network application

Computing in memory (CIM) based on memristors is expected to completely solve the dilemma caused by von Neumann architecture. However, the performance of memristors based on traditional conductive filament mechanism is unstable. In this study, we report a nonvolatile high-performance memristor based on ferroelectric tunnel junction (FTJ) Pd/Bi0.9La0.1FeO3 (6.9 nm) (BLFO)/La0.67Sr0.33MnO3 (LSMO) on a silicon substrate. The conductance of this device was adjusted by different pulse stimulation parameter to achieve various synaptic functions because of ferroelectric polarization reversal. Based on the multiple conductance characteristics of the devices and the high linearity and symmetry of weight updating, image processing and VGG8 convolutional neural network (CNN) simulation based on the devices were realized. Excellent results of the image processing are demonstrated. The recognition accuracy of CNN offline learning reached an astonishing 92.07% based on Cifar-10 dataset. This provides a more feasible solution to break through the bottleneck of von Neumann architecture.

Integration of image preprocessing and recognition functions in an optoelectronic coupling organic ferroelectric retinomorphic neuristor

The human visual system (HVS) has the advantages of a low power consumption and high efficiency because of the synchronous perception and early preprocessing of external image information in the retina, as well as parallel in-memory computing within the visual cortex. Realizing the biofunction simulation of the retina and visual cortex in a single device structure provides opportunities for performance improvements and machine vision system (MVS) integration. Here, we fabricate organic ferroelectric retinomorphic neuristors that integrate the retina-like preprocessing function and recognition of the visual cortex in a single device architecture. Benefiting from the electrical/optical coupling modulation of ferroelectric polarization, our devices show a bidirectional photoresponse that acts as the basis for mimicking retinal preconditioning and multi-level memory capabilities for recognition. The MVS based on the proposed retinomorphic neuristors achieves a high recognition accuracy of ∼90%, which is 20% higher than that of the incomplete system without the preprocessing function. In addition, we successfully demonstrate image encryption and optical programming logic gate functions. Our work suggests that the proposed retinomorphic neuristors offer great potential for MVS monolithic integration and functional expansion.

All-ferroelectric implementation of reservoir computing

Reservoir computing (RC) offers efficient temporal information processing with low training cost. All-ferroelectric implementation of RC is appealing because it can fully exploit the merits of ferroelectric memristors (e.g., good controllability); however, this has been undemonstrated due to the challenge of developing ferroelectric memristors with distinctly different switching characteristics specific to the reservoir and readout network. Here, we experimentally demonstrate an all-ferroelectric RC system whose reservoir and readout network are implemented with volatile and nonvolatile ferroelectric diodes (FDs), respectively. The volatile and nonvolatile FDs are derived from the same Pt/BiFeO₃/SrRuO₃ structure via the manipulation of an imprint field (E_imp). It is shown that the volatile FD with E_imp exhibits short-term memory and nonlinearity while the nonvolatile FD with negligible E_imp displays long-term potentiation/depression, fulfilling the functional requirements of the reservoir and readout network, respectively. Hence, the all-ferroelectric RC system is competent for handling various temporal tasks. In particular, it achieves an ultralow normalized root mean square error of 0.017 in the Hénon map time-series prediction. Besides, both the volatile and nonvolatile FDs demonstrate long-term stability in ambient air, high endurance, and low power consumption, promising the all-ferroelectric RC system as a reliable and low-power neuromorphic hardware for temporal information processing.

Ultralow-power in-memory computing based on ferroelectric memcapacitor network

Analog storage through synaptic weights using conductance in resistive neuromorphic systems and devices inevitably generates harmful heat dissipation. This thermal issue not only limits the energy efficiency but also hampers the very-large-scale and highly complicated hardware integration as in the human brain. Here we demonstrate that the synaptic weights can be simulated by reconfigurable non-volatile capacitances of a ferroelectric-based memcapacitor with ultralow-power consumption. The as-designed metal/ferroelectric/metal/insulator/semiconductor memcapacitor shows distinct 3-bit capacitance states controlled by the ferroelectric domain dynamics. These robust memcapacitive states exhibit uniform maintenance of more than 104 s and well endurance of 109 cycles. In a wired memcapacitor crossbar network hardware, analog vector-matrix multiplication is successfully implemented to classify 9-pixel images by collecting the sum of displacement currents (I = C × dV/dt) in each column, which intrinsically consumes zero energy in memcapacitors themselves. Our work sheds light on an ultralow-power neural hardware based on ferroelectric memcapacitors.

Understanding the Effect of Top Electrode on Ferroelectricity in Atomic Layer Deposited Hf0.5Zr0.5O2 Thin Films

It is commonly believed that the impact of the top electrodes on the ferroelectricity of hafnium-based thin films is due to strain engineering. However, several anomalies have occurred that put existing theories in doubt. This work carries out a detailed study of this issue using both theoretical and experimental approaches. The 10 nm Hf_0.5Zr_0.5O₂ (HZO) films are prepared by atomic layer deposition, and three different top capping electrodes (W/MO/ITO) are deposited by physical vapor deposition. The electrical testing finds that the strain does not completely control the ferroelectricity of the devices. The results of further piezoelectric force microscopy characterization exclude the potential interference of the top capping electrodes and interface for electrical testing. In addition, through atomic force microscopy characterization and statistical analysis, a strong correlation between the grain size of the top electrode and the grain size of the HZO film has been found, suggesting that the grain size of the top electrode can induce the formation of the grain size in HZO thin films. Finally, the first-principles calculation is carried out to understand the impact of the strain and grain size on the ferroelectric properties of HZO films. The results show that the strain is the dominant factor for ferroelectricity when the grain size is large (>10 nm). However, when the grain size becomes thinner (<10 nm), the regulation effect of grain sizes increases significantly, which could bring a series of benefits for device scaling, such as device-to-device variations, film uniformity, and domain switch consistency. This work not only completes the understanding of ferroelectricity through top electrode modulation but also provides strong support for the precise regulation of ferroelectricity of nanoscale devices and ultrathin HZO ferroelectric films.

Intelligent matter endows reconfigurable temperature and humidity sensations for in-sensor computing

Data-centric tactics with in-sensor computing go beyond the conventional computing-centric tactic that is suffering from processing latency and excessive energy consumption. The multifunctional intelligent matter with dynamic smart responses to environmental variations paves the way to implement data-centric tactics with high computing efficiency. However, intelligent matter with humidity and temperature sensitivity has not been reported. In this work, a design is demonstrated based on a single memristive device to achieve reconfigurable temperature and humidity sensations. Opposite temperature sensations at the low resistance state (LRS) and high resistance state (HRS) were observed for low-level sensory data processing. Integrated devices mimicking intelligent electronic skin (e-skin) can work in three modes to adapt to different scenarios. Additionally, the device acts as a humidity-sensory artificial synapse that can implement high-level cognitive in-sensor computing. The intelligent matter with reconfigurable temperature and humidity sensations is promising for energy-efficient artificial intelligence (AI) systems.

A Ferroelectric Memristor-Based Transient Chaotic Neural Network for Solving Combinatorial Optimization Problems

A transient chaotic neural network (TCNN) is particularly useful for solving combinatorial optimization problems, and its hardware implementation based on memristors has attracted great attention recently. Although previously used filamentary memristors could provide the desired nonlinearity for implementing the annealing function of a TCNN, the controllability of filamentary switching still remains relatively poor, thus limiting the performance of a memristor-based TCNN. Here, we propose to use ferroelectric memristor to implement the annealing function of a TCNN. In the ferroelectric memristor, the conductance can be tuned by switching the lattice non-centrosymmetry-induced polarization, which is a nonlinear switching mechanism with high controllability. We first establish a ferroelectric memristor model based on a ferroelectric tunnel junction (FTJ), which exhibits the polarization-modulated tunnel conductance and the nucleation-limited-switching (NLS) behavior. Then, the conductance of the ferroelectric memristor is used as the self-feedback connection weight that can be dynamically adjusted. Based on this, a ferroelectric memristor-based transient chaotic neural network (FM-TCNN) is further constructed and applied to solve the traveling salesman problem (TSP). In 1000 runs for 10-city TSP, the FM-TCNN achieves a shorter average path distance, a 32.8% faster convergence speed, and a 2.44% higher global optimal rate than the TCNN.