Resistive random-access memory(RRAM),also known as memristors,having a very simple device structure with two terminals,fulfill almost all of the fundamental requirements of volatile memory,nonvolatile memory,and neuro...Resistive random-access memory(RRAM),also known as memristors,having a very simple device structure with two terminals,fulfill almost all of the fundamental requirements of volatile memory,nonvolatile memory,and neuromorphic characteristics.Its memory and neuromorphic behaviors are currently being explored in relation to a range of materials,such as biological materials,perovskites,2D materials,and transition metal oxides.In this review,we discuss the different electrical behaviors exhibited by RRAM devices based on these materials by briefly explaining their corresponding switching mechanisms.We then discuss emergent memory technologies using memristors,together with its potential neuromorphic applications,by elucidating the different material engineering techniques used during device fabrication to improve the memory and neuromorphic performance of devices,in areas such as ION/IOFF ratio,endurance,spike time-dependent plasticity(STDP),and paired-pulse facilitation(PPF),among others.The emulation of essential biological synaptic functions realized in various switching materials,including inorganic metal oxides and new organic materials,as well as diverse device structures such as single-layer and multilayer hetero-structured devices,and crossbar arrays,is analyzed in detail.Finally,we discuss current challenges and future prospects for the development of inorganic and new materials-based memristors.展开更多
Memristors are now becoming a prominent candidate to serve as the building blocks of non-von Neumann inmemory computing architectures.By mapping analog numerical matrices into memristor crossbar arrays,efficient multi...Memristors are now becoming a prominent candidate to serve as the building blocks of non-von Neumann inmemory computing architectures.By mapping analog numerical matrices into memristor crossbar arrays,efficient multiply accumulate operations can be performed in a massively parallel fashion using the physics mechanisms of Ohm’s law and Kirchhoff’s law.In this brief review,we present the recent progress in two niche applications:neural network accelerators and numerical computing units,mainly focusing on the advances in hardware demonstrations.The former one is regarded as soft computing since it can tolerant some degree of the device and array imperfections.The acceleration of multiple layer perceptrons,convolutional neural networks,generative adversarial networks,and long short-term memory neural networks are described.The latter one is hard computing because the solving of numerical problems requires high-precision devices.Several breakthroughs in memristive equation solvers with improved computation accuracies are highlighted.Besides,other nonvolatile devices with the capability of analog computing are also briefly introduced.Finally,we conclude the review with discussions on the challenges and opportunities for future research toward realizing memristive analog computing machines.展开更多
The development of electronic devices that possess the functionality of biological synapses is a crucial step towards neuromorphic computing.In this work,we present a WOx-based memristive device that can emulate volta...The development of electronic devices that possess the functionality of biological synapses is a crucial step towards neuromorphic computing.In this work,we present a WOx-based memristive device that can emulate voltage-dependent synaptic plasticity.By adjusting the amplitude of the applied voltage,we were able to reproduce short-term plasticity(STP)and the transition from STP to long-term potentiation.The stimulation with high intensity induced long-term enhancement of conductance without any decay process,thus representing a permanent memory behavior.Moreover,the image Boolean operations(including intersection,subtraction,and union)were also demonstrated in the memristive synapse array based on the above voltage-dependent plasticity.The experimental achievements of this study provide a new insight into the successful mimicry of essential characteristics of synaptic behaviors.展开更多
Flexible resistive random access memory(RRAM) has shown great potential in wearable electronics.With tunable multilevel resistance states,flexible memristors could be used to mimic the bio-synapses for constructing hi...Flexible resistive random access memory(RRAM) has shown great potential in wearable electronics.With tunable multilevel resistance states,flexible memristors could be used to mimic the bio-synapses for constructing high-efficient wearable neuromorphic computing system.However,the flexible substrate has intrinsic disadvantages including low-tempe rature tolerance and poor complementary metal-oxidesemiconductor(CMOS) compatibility,which limit the development of flexible electronics.The physical vapor deposition(PVD) fabrication process could prepare RRAM without requirement of further treatment,which greatly simplified preparation steps and reduced the production costs.On the other hand,forming process,as a common pre-programing operation in RRAM,increases the energy consumption and limits the application scenarios of RRAM.Here,a NiO-based forming-free RRAM with low set voltage was fabricated via full PVD technique.The flexible device exhibited reliable re sistive switching characteristics under flat state even compre s sive and tensile states(R=10 mm).The tunable multilevel resistance states(5 levels) could be obtained by controlling the compliance current.Besides,synaptic plasticities also were verified in this device.The flexible NiO-based RRAM shows great potential in wearable forming-free multibit memo ry and neuromorphic computing electronics.展开更多
In this paper, a large dynamic range floating memristor emulator(LDRFME) with equal port current restriction is proposed to be achieved by a large dynamic range floating voltage-controlled linear resistor(VCLR). Since...In this paper, a large dynamic range floating memristor emulator(LDRFME) with equal port current restriction is proposed to be achieved by a large dynamic range floating voltage-controlled linear resistor(VCLR). Since real memristors have not been largely commercialized until now, the application of a LDRFME to memristive systems is reasonable. Motivated by this need, this paper proposes an achievement of a LDRFME based on a feasible transistor model. A first circuit extends the voltage range of the triode region of an ordinary junction field effect transistor(JFET). The idea is to use this JFET transistor as a tunable linear resistor. A second memristive non-linear circuit is used to drive the resistance of the first JFET transistor. Then those two circuits are connected together and, under certain conditions, the obtained "resistor" presents a hysteretic behavior,which is considered as a memristive effect. The electrical characteristics of a LDRFME are validated by software simulation and real measurement, respectively.展开更多
Image fusion is widely used in computer vision and image analysis.Considering that the traditional image fusion algorithm has a certain limitation in multi-channel image fusion,a memristor-based multi-channel pulse co...Image fusion is widely used in computer vision and image analysis.Considering that the traditional image fusion algorithm has a certain limitation in multi-channel image fusion,a memristor-based multi-channel pulse coupled neural network(M-MPCNN)for image fusion is proposed.Based on a dual-channel pulse coupled neural network(D-PCNN),a novel multi-channel pulse coupled neural network(M-PCNN)is firstly constructed in this paper.Then the exponential growth dynamic threshold model is used to improve the pulse generation of pulse coupled neural network,which can not only avoid multiple ignitions effectively,but can also improve operational efficiency and reduce complexity.At the same time,synchronous capture can also enhance image edge,which is more conducive to image fusion.Finally,the threshold and synaptic characteristics of pulse coupled neural networks(PCNNs)can be well realized by using a memristor-based pulse generator.Experimental results show that the proposed algorithm can fuse multi-source images more effectively than existing state-of-the-art fusion algorithms.展开更多
Complexity and abundant dynamics may arise in locally-active systems only, in which locally-active elements are essential to amplify infinitesimal fluctuation signals and maintain oscillating. It has been recently fou...Complexity and abundant dynamics may arise in locally-active systems only, in which locally-active elements are essential to amplify infinitesimal fluctuation signals and maintain oscillating. It has been recently found that some memristors may act as locally-active elements under suitable biasing. A number of important engineering applications would benefit from locally-active memristors. The aim of this paper is to show that locally-active memristor-based circuits can generate periodic and chaotic oscillations. To this end, we propose a non-volatile locally-active memristor, which has two asymptotically stable equilibrium points(or two non-volatile memristances) and globally-passive but locally-active characteristic. At an operating point in the locally-active region, a small-signal equivalent circuit is derived for describing the characteristics of the memristor near the operating point. By using the small-signal equivalent circuit, we show that the memristor possesses an edge of chaos in a voltage range, and that the memristor, when connected in series with an inductor,can oscillate about a locally-active operating point in the edge of chaos. And the oscillating frequency and the external inductance are determined by the small-signal admittance Y(iω). Furthermore, if the parasitic capacitor in parallel with the memristor is considered in the periodic oscillating circuit, the circuit generates chaotic oscillations.展开更多
基金Basic Science Research Program through the National Research Foundation of Korea(NRF),funded by the Ministry of Education(NRF-2019R1F1A1057243)together with the Future Semiconductor Device Technology Development Program(20003808,10080689,20004399)funded by MOTIE(Ministry of Trade,Industry&Energy)and KSRC(Korea Semiconductor Research Consortium).
文摘Resistive random-access memory(RRAM),also known as memristors,having a very simple device structure with two terminals,fulfill almost all of the fundamental requirements of volatile memory,nonvolatile memory,and neuromorphic characteristics.Its memory and neuromorphic behaviors are currently being explored in relation to a range of materials,such as biological materials,perovskites,2D materials,and transition metal oxides.In this review,we discuss the different electrical behaviors exhibited by RRAM devices based on these materials by briefly explaining their corresponding switching mechanisms.We then discuss emergent memory technologies using memristors,together with its potential neuromorphic applications,by elucidating the different material engineering techniques used during device fabrication to improve the memory and neuromorphic performance of devices,in areas such as ION/IOFF ratio,endurance,spike time-dependent plasticity(STDP),and paired-pulse facilitation(PPF),among others.The emulation of essential biological synaptic functions realized in various switching materials,including inorganic metal oxides and new organic materials,as well as diverse device structures such as single-layer and multilayer hetero-structured devices,and crossbar arrays,is analyzed in detail.Finally,we discuss current challenges and future prospects for the development of inorganic and new materials-based memristors.
基金supported by the National Natural Science Foundation of China(61674050 and 61874158)the Outstanding Youth Funding of Hebei University(A2018201231)+2 种基金the Support Program for the Top Young Talents of Hebei Province(70280011807)the Hundred Persons Plan of Hebei Province(E2018050004 and E2018050003)the Supporting Plan for 100 Excellent Innovative Talents in Colleges and Universities of Hebei Province(SLRC2019018)。
基金the National Key Research and Development Plan of MOST of China(2019YFB2205100,2016YFA0203800)the National Natural Science Foundation of China(No.61874164,61841404,51732003,61674061)Hubei Engineering Research Center on Microelectronics.
文摘Memristors are now becoming a prominent candidate to serve as the building blocks of non-von Neumann inmemory computing architectures.By mapping analog numerical matrices into memristor crossbar arrays,efficient multiply accumulate operations can be performed in a massively parallel fashion using the physics mechanisms of Ohm’s law and Kirchhoff’s law.In this brief review,we present the recent progress in two niche applications:neural network accelerators and numerical computing units,mainly focusing on the advances in hardware demonstrations.The former one is regarded as soft computing since it can tolerant some degree of the device and array imperfections.The acceleration of multiple layer perceptrons,convolutional neural networks,generative adversarial networks,and long short-term memory neural networks are described.The latter one is hard computing because the solving of numerical problems requires high-precision devices.Several breakthroughs in memristive equation solvers with improved computation accuracies are highlighted.Besides,other nonvolatile devices with the capability of analog computing are also briefly introduced.Finally,we conclude the review with discussions on the challenges and opportunities for future research toward realizing memristive analog computing machines.
基金the fund from Ministry of Science and Technology of China(Nos.2018YFE0118300 and 2019YFB2205100)the NSFC Program(Nos.11974072,51701037,51732003,51872043,51902048,61774031,61574031 and U19A2091)+4 种基金the“111”Project(No.B13013)the fund from Ministry of Education of China(No.6141A02033414)The fund from China Postdoctoral Science Foundation(No.2019M661185)The Fundamental Research Funds for the Central Universities(No.2412019QD015)the Fund from Jilin Province(JJKH20201163KJ).
文摘The development of electronic devices that possess the functionality of biological synapses is a crucial step towards neuromorphic computing.In this work,we present a WOx-based memristive device that can emulate voltage-dependent synaptic plasticity.By adjusting the amplitude of the applied voltage,we were able to reproduce short-term plasticity(STP)and the transition from STP to long-term potentiation.The stimulation with high intensity induced long-term enhancement of conductance without any decay process,thus representing a permanent memory behavior.Moreover,the image Boolean operations(including intersection,subtraction,and union)were also demonstrated in the memristive synapse array based on the above voltage-dependent plasticity.The experimental achievements of this study provide a new insight into the successful mimicry of essential characteristics of synaptic behaviors.
基金supported by the National Natural Science Foundation of China(Nos.61704030 and 61522404)the Shanghai Rising-Star Program(No.19QA1400600)+1 种基金the Program of Shanghai Subject Chief Scientist(No.18XD1402800)the Support Plans for the Youth Top-Notch Talents of China。
文摘Flexible resistive random access memory(RRAM) has shown great potential in wearable electronics.With tunable multilevel resistance states,flexible memristors could be used to mimic the bio-synapses for constructing high-efficient wearable neuromorphic computing system.However,the flexible substrate has intrinsic disadvantages including low-tempe rature tolerance and poor complementary metal-oxidesemiconductor(CMOS) compatibility,which limit the development of flexible electronics.The physical vapor deposition(PVD) fabrication process could prepare RRAM without requirement of further treatment,which greatly simplified preparation steps and reduced the production costs.On the other hand,forming process,as a common pre-programing operation in RRAM,increases the energy consumption and limits the application scenarios of RRAM.Here,a NiO-based forming-free RRAM with low set voltage was fabricated via full PVD technique.The flexible device exhibited reliable re sistive switching characteristics under flat state even compre s sive and tensile states(R=10 mm).The tunable multilevel resistance states(5 levels) could be obtained by controlling the compliance current.Besides,synaptic plasticities also were verified in this device.The flexible NiO-based RRAM shows great potential in wearable forming-free multibit memo ry and neuromorphic computing electronics.
基金the National Key Research and Development Program of China(2018YFC0830300)the National Natural Science Foundation of China(61571312)the Science and Technology Support Project of Chengdu PU Chip Science and Technology Co.,Ltd.
文摘In this paper, a large dynamic range floating memristor emulator(LDRFME) with equal port current restriction is proposed to be achieved by a large dynamic range floating voltage-controlled linear resistor(VCLR). Since real memristors have not been largely commercialized until now, the application of a LDRFME to memristive systems is reasonable. Motivated by this need, this paper proposes an achievement of a LDRFME based on a feasible transistor model. A first circuit extends the voltage range of the triode region of an ordinary junction field effect transistor(JFET). The idea is to use this JFET transistor as a tunable linear resistor. A second memristive non-linear circuit is used to drive the resistance of the first JFET transistor. Then those two circuits are connected together and, under certain conditions, the obtained "resistor" presents a hysteretic behavior,which is considered as a memristive effect. The electrical characteristics of a LDRFME are validated by software simulation and real measurement, respectively.
基金This work was supported by the National Natural Science Foundation of China(61671377,51709228)the Shaanxi Natural Science Foundation of China(2016JM8034,2017JM6107).
文摘Image fusion is widely used in computer vision and image analysis.Considering that the traditional image fusion algorithm has a certain limitation in multi-channel image fusion,a memristor-based multi-channel pulse coupled neural network(M-MPCNN)for image fusion is proposed.Based on a dual-channel pulse coupled neural network(D-PCNN),a novel multi-channel pulse coupled neural network(M-PCNN)is firstly constructed in this paper.Then the exponential growth dynamic threshold model is used to improve the pulse generation of pulse coupled neural network,which can not only avoid multiple ignitions effectively,but can also improve operational efficiency and reduce complexity.At the same time,synchronous capture can also enhance image edge,which is more conducive to image fusion.Finally,the threshold and synaptic characteristics of pulse coupled neural networks(PCNNs)can be well realized by using a memristor-based pulse generator.Experimental results show that the proposed algorithm can fuse multi-source images more effectively than existing state-of-the-art fusion algorithms.
基金Project supported by the National Natural Science Foundation of China(Grant No.61771176)。
文摘Complexity and abundant dynamics may arise in locally-active systems only, in which locally-active elements are essential to amplify infinitesimal fluctuation signals and maintain oscillating. It has been recently found that some memristors may act as locally-active elements under suitable biasing. A number of important engineering applications would benefit from locally-active memristors. The aim of this paper is to show that locally-active memristor-based circuits can generate periodic and chaotic oscillations. To this end, we propose a non-volatile locally-active memristor, which has two asymptotically stable equilibrium points(or two non-volatile memristances) and globally-passive but locally-active characteristic. At an operating point in the locally-active region, a small-signal equivalent circuit is derived for describing the characteristics of the memristor near the operating point. By using the small-signal equivalent circuit, we show that the memristor possesses an edge of chaos in a voltage range, and that the memristor, when connected in series with an inductor,can oscillate about a locally-active operating point in the edge of chaos. And the oscillating frequency and the external inductance are determined by the small-signal admittance Y(iω). Furthermore, if the parasitic capacitor in parallel with the memristor is considered in the periodic oscillating circuit, the circuit generates chaotic oscillations.