Article
Multidisciplinary Sciences
Piotr Rzeszut, Jakub Checinski, Ireneusz Brzozowski, Slawomir Zietek, Witold Skowronski, Tomasz Stobiecki
Summary: In this study, a complete hardware implementation design of a neural computing device that incorporates serially connected MTJs forming a multi-state memory cell is presented. The designed network shows a comparable detection ratio to the software algorithm in recognizing hand-written digits, using weights stored in a multi-cell consisting of four or more MTJs. Moreover, the presented solution has better energy efficiency in terms of energy consumed per single image processing, compared to a similar design.
SCIENTIFIC REPORTS
(2022)
Article
Computer Science, Artificial Intelligence
Xiaoyang Liu, Zhigang Zeng
Summary: This paper introduces memristor crossbar architectures for implementing various layers in deep neural networks, and analyzes the impact of the inherent characteristics of memristors and programming voltage errors on the networks. Simulation results show that deep neural networks built by memristor crossbars perform well in pattern recognition tasks.
COMPLEX & INTELLIGENT SYSTEMS
(2022)
Article
Multidisciplinary Sciences
Rui Wang, Tuo Shi, Xumeng Zhang, Jinsong Wei, Jian Lu, Jiaxue Zhu, Zuheng Wu, Qi Liu, Ming Liu
Summary: This article introduces a hardware implementation of self-organizing maps (SOM) based on memristors, which provides faster computing speed and energy efficiency compared to CMOS digital counterparts. The memristor-based SOM demonstrates improved performance in data clustering, image processing, and optimization problem-solving.
NATURE COMMUNICATIONS
(2022)
Article
Automation & Control Systems
Changbao Wen, Jun Zha, Li Xu, Feng Ru, Si Quan
Summary: This article proposes a design scheme of perceptron neural networks based on memristors, and verifies its feasibility through experiments on multiclassification and linear indivisibility problems.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Mathematics, Interdisciplinary Applications
Shoukui Ding, Ning Wang, Han Bao, Bei Chen, Huagan Wu, Quan Xu
Summary: This paper proposes a new neural network model based on memristors to simulate the electromagnetic induction effect between neurons. The theoretical analysis and numerical simulations investigate the multistability and various dynamic behaviors of the model, and a simple analog circuit is designed for verification.
CHAOS SOLITONS & FRACTALS
(2023)
Review
Physics, Multidisciplinary
Longcheng Ye, Zhixuan Gao, Jinke Fu, Wang Ren, Cihui Yang, Jing Wen, Xiang Wan, Qingying Ren, Shipu Gu, Xiaoyan Liu, Xiaojuan Lian, Lei Wang
Summary: Conventional von Newmann-based computers face challenges in processing and storing big data, which can be addressed by using artificial neural networks (ANN). By using memristors instead of traditional components, the limitations of Moore's law can be extended. This paper classifies and discusses the physical principles, technologies, and applications of memristors, and envisions their potential in neural networks.
FRONTIERS IN PHYSICS
(2022)
Article
Computer Science, Artificial Intelligence
Shaobo He, Jun Liu, Huihai Wang, Kehui Sun
Summary: The design of artificial neural networks based on memristors has received increasing attention from researchers. However, there have been no reports of discrete memristor-based neural networks. In this study, a novel discrete memristor BP neural network is designed. A three-layer discrete memristor BP neural network is built for MNIST-10 handwriting recognition, achieving a high classification accuracy of 97.16%, which demonstrates the effectiveness of the proposed method.
Article
Physics, Multidisciplinary
Kai-Da Xu, Donghao Li, Yannan Jiang, Qiang Chen
Summary: This paper presents a new HP memristor model with a new window function, and studies the SPICE behaviors of the linear and nonlinear memristor model through PSpice simulation. It demonstrates high flexibility in emulating the practical HP memristors and investigates the characteristics of composite SPICE behaviors in series and parallel connections of two memristors. The relationships among flux, charge, voltage, current, and memristance in double memristor circuits are simulated and analyzed.
FRONTIERS IN PHYSICS
(2021)
Article
Chemistry, Multidisciplinary
Shuai Fu, Ji-Hoon Park, Hongyan Gao, Tianyi Zhang, Xiang Ji, Tianda Fu, Lu Sun, Jing Kong, Jun Yao
Summary: Research demonstrates the construction of programmable networks using two-terminal MoS2 memristors. These memristors work with a charge-based mechanism similar to transistors, allowing for homogeneous integration with MoS2 transistors to realize one-transistor-one-memristor addressable cells.
Article
Chemistry, Multidisciplinary
Zilong Dong, Qilin Hua, Jianguo Xi, Yuanhong Shi, Tianci Huang, Xinhuan Dai, Jianan Niu, Bingjun Wang, Zhong Lin Wang, Weiguo Hu
Summary: Memristors that mimic synaptic plasticity are crucial for energy-efficient neuromorphic computing architecture, and layered 2D Bi2O2Se is an important material in improving memristive device efficiency. High-quality Bi2O2Se nanosheets are grown on mica substrates, and bipolar Bi2O2Se memristors with outstanding performance are fabricated. These memristors exhibit ultrafast switching speed (<5 ns), low power consumption (<3.02 pJ), and demonstrate synaptic plasticity. Utilizing conductance modification in simulated artificial neural networks (ANN), MNIST recognition achieves high accuracy of 91%. The 2D Bi2O2Se enables the memristors to possess ultrafast and low-power attributes, showing great potential in neuromorphic computing applications.
Article
Engineering, Electrical & Electronic
Danzhe Song, Fan Yang, Chengxu Wang, Nan Li, Pinfeng Jiang, Bin Gao, Xiangshui Miao, Xingsheng Wang
Summary: In this work, two innovative schemes from the level of software are proposed to mitigate the hardware IR-drop problem caused by line resistance in a large-scale memristor crossbar array. The methods are tested on MLP and LeNet-5 neural networks for MNIST recognition using typical activation functions and various line resistances. Results demonstrate that the methods can significantly improve the tolerance of neural networks to IR-drop and recover accuracy to some extent. These methods require no additional hardware overhead and reduce the complexity of peripheral circuits, making them more achievable and attractive.
IEEE ELECTRON DEVICE LETTERS
(2023)
Article
Mathematics, Interdisciplinary Applications
Wei Zhou, Bowei Fu, Guangyi Wang
Summary: This paper focuses on the hybrid effects of memristor characteristics, coupling coefficient, and time-delay on the recurrent neural network. A novel time-delay recurrent neural network based on a passive hyperbolic tangent memristor is constructed, and its dynamic behaviors are analyzed. The theoretical results are verified through circuit simulation.
INTERNATIONAL JOURNAL OF BIFURCATION AND CHAOS
(2022)
Article
Engineering, Electrical & Electronic
Mubeen Zafar, Muhammad Naeem Awais, Muhammad Naeem Shehzad
Summary: This research develops a novel memristor model using a Hann window function to address the limitations of existing mathematical models. The proposed model has a wider frequency range and voltage range compared to existing models. By incorporating a scaling parameter, the model's scalability and flexibility are improved. The proposed model demonstrates improved simulation runtime and accuracy.
MICROELECTRONICS JOURNAL
(2022)
Article
Computer Science, Artificial Intelligence
Haijun Lin, Houde Dai, Yihan Mao, Lucai Wang
Summary: This paper proposes a novel MW-RBFNN model with an adjustable activation function, which can improve the performance of neural networks. Compared with the traditional RC-RBF neural network, MW-RBFNN has stronger approximating ability and lower computational complexity.
Article
Mathematics, Interdisciplinary Applications
Shasha Wang, Jigui Jian
Summary: This article investigates the predefined-time synchronization of fractional-order memristive competitive neural networks with time-varying delays. Two distinct discontinuous bilayer predefined-time control schemes based on fractional integrals are proposed to address the two-layer structural characteristics of CNNs. By utilizing predefined-time stability theorems and applying fractional-order differential inequalities and other inequality techniques, concise criteria are obtained to ensure the predefined-time stability of two FMCNNs in terms of algebraic inequalities. The predefined time parameter is arbitrary and does not depend on the initial values. Two examples are provided to validate the theoretical results.
CHAOS SOLITONS & FRACTALS
(2023)
Article
Computer Science, Artificial Intelligence
Zi-Peng Wang, Qian-Qian Li, Huai-Ning Wu, Biao Luo, Tingwen Huang
Summary: This article investigates the pinning spatiotemporal sampled-data synchronization problem of coupled reaction-diffusion neural networks under random deception attacks. A directed CRDNN model is established to handle the impacts of variable sampling and random deception attacks within a unified framework. Sufficient conditions are obtained through a designed pinning spatiotemporal SD controller, ensuring the stability of the synchronization error system.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Chao-Yang Chen, Yong Zhou, Yanwu Wang, Li Ding, Tingwen Huang
Summary: In this article, a new electrical betweenness method is proposed to identify vulnerable lines in the power grid, which can balance the accuracy and efficiency. The power percentage of generator-load is used as an indicator to evaluate the stability of the power grid, considering the dynamic influence of generator and load node removal. The nonlinear cascading failure model with overload and weighted edges is introduced, considering the three states of normal-failure-overload of lines, making it more in line with the actual power system. Simulation analysis on IEEE 39 and 118-bus system verifies the feasibility and effectiveness of this method in identifying vulnerable lines in the cascading failure model by comparing the changes in power loss caused by removing a certain proportion of vulnerable lines and deliberately attacking high new electrical betweenness.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS
(2023)
Article
Computer Science, Artificial Intelligence
Yuting Cao, Linhao Zhao, Qishui Zhong, Shiping Wen, Kaibo Shi, Jianying Xiao, Tingwen Huang
Summary: This paper addresses the fixed-time output synchronization problems for two types of complex dynamical networks with multi-weights (CDNMWs) by using two types of adaptive control methods. The networks with multiple state and output couplings are presented, and fixed-time output synchronization criteria are formulated based on Lyapunov functional and inequality techniques. The fixed-time output synchronization issues of these networks are dealt with using the two types of adaptive control methods. The analytical results are verified by two numerical simulations.
Article
Automation & Control Systems
Dongxue Jiang, Guoguang Wen, Zhaoxia Peng, Jin-Liang Wang, Tingwen Huang
Summary: This article addresses the fixed-time output consensus problem of heterogeneous linear multiagent systems with an active leader. It proposes a fully distributed fixed-time observer and adaptive algorithms to estimate the leader's system matrices and solve the regulator equations. By employing the pull-based event-triggered mechanism, it introduces observer-based controllers and provides sufficient criteria to ensure the bipartite output consensus in fixed time, ruling out Zeno behavior.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Automation & Control Systems
Shengbo Wang, Bo Lyu, Shiping Wen, Kaibo Shi, Song Zhu, Tingwen Huang
Summary: This article investigates a safety-critical control scheme for unknown structured systems using the control barrier function (CBF) method and dynamic regressor extension and mixing (DREM) method. The proposed scheme ensures safety in the identification process and minimizes theoretical conservatism compared to other existing adaptive CBF algorithms. The stability and robustness of the scheme under bounded disturbances are analyzed, and simulation-based examples demonstrate its effectiveness.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Editorial Material
Computer Science, Cybernetics
Tingwen Huang
IEEE SYSTEMS MAN AND CYBERNETICS MAGAZINE
(2023)
Article
Automation & Control Systems
Shengbo Wang, Shiping Wen, Yin Yang, Yuting Cao, Kaibo Shi, Tingwen Huang
Summary: This article presents an innovative efficient safety-critical control scheme for nonlinear systems by combining techniques of control barrier function (CBF) and online time-varying optimization. The effectiveness of the scheme is demonstrated through two experiments on obstacle avoidance and anti-swing tasks.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Computer Science, Information Systems
Ziyu Sheng, Shiping Wen, Zhong-Kai Feng, Kaibo Shi, Tingwen Huang
Summary: Runoff forecasting is crucial for the rational use and protection of water resources. This article proposes a novel framework called ResGRU Plus, which combines GRU, ResNet, and SENet to improve the depth and accuracy of the model. Multiple experiments show that ResGRU Plus outperforms traditional models and achieves state-of-the-art performance in runoff forecasting.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Automation & Control Systems
Liqing Wang, Zheng-Guang Wu, Tingwen Huang, Prasun Chakrabarti, Wei-Wei Che
Summary: This article studies the finite-time observability of switching Boolean networks with Markov jump parameters. By using a parallel extension method, the observability of the considered networks is equivalent to the reachability of the zero vector from an initial set of the newly constructed system. A necessary and sufficient condition based on the extended structure matrix is proposed for finite-time observability. Furthermore, mode-dependent pinning control is introduced and applied for unobservable systems to achieve observability.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Automation & Control Systems
Hongsong Wen, Xing He, Tingwen Huang, Junzhi Yu
Summary: This article develops several neurodynamic algorithms for sparse signal recovery by solving the l(1) regularization problem. The solutions of the proposed algorithms exist and are unique under the observation matrix satisfying restricted isometry property (RIP) condition, while their convergence to the optimal points is shown via Lyapunov-based analysis. Upper bounds on the convergence time of the algorithms are given, and the convergence results obtained for certain algorithms are shown to be independent of the initial conditions. Simulation experiments demonstrate the superior performance of the proposed algorithms for signal recovery and image recovery.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Qingguo Lu, Shaojiang Deng, Huaqing Li, Tingwen Huang
Summary: Decentralized algorithms for economic dispatch problem in smart grids have gained significant attention due to their scalability, robustness, and flexibility. We propose a novel privacy-protected decentralized random sleep algorithm to address issues of data security and computation efficiency, and provide theoretical proof of its optimality, convergence, and privacy properties.
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Hailong Tan, Bo Shen, Qi Lid, Tingwen Huang
Summary: This paper studies the problem of zonotopic set-membership estimation (SME) for time-varying systems subject to dynamical biases and uniform quantization. A mathematical method is proposed to estimate the state of the system by analyzing the dynamics of biases and states. An auxiliary zonotope is constructed to minimize the estimation error, and an external approximation is used to reduce the computational burden. The effectiveness of the proposed method is demonstrated through simulations.
INFORMATION SCIENCES
(2024)
Article
Computer Science, Artificial Intelligence
Xin Han, Xing He, Xingxing Ju, Hangjun Che, Tingwen Huang
Summary: This article investigates a class of systems of nonlinear equations and proposes three distributed neurodynamic models for solving such equations. These models have global convergence and certain advantages, are applicable to both smooth and nonsmooth cases, and have good performance in quadratic programming problems.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Automation & Control Systems
Yuan Yuan, Xiaodong Xu, Chunhua Yang, Tingwen Huang, Stevan Dubljevic
Summary: In this paper, a solution for fault estimation is proposed for one-dimensional linear boundary control and boundary observation parabolic PDEs, which can accurately estimate the unknown multiplicative fault parameter in the measurement.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Depeng Li, Tianqi Wang, Junwei Chen, Kenji Kawaguchi, Cheng Lian, Zhigang Zeng
Summary: This paper investigates a novel paradigm called multi-view class incremental learning (MVCIL), which addresses the challenges of catastrophic forgetting and interference in multi-view learning. The paper proposes a randomization-based representation learning technique and selective weight consolidation to tackle these challenges. Extensive experiments validate the effectiveness of the approach.
INFORMATION FUSION
(2024)
Article
Computer Science, Artificial Intelligence
Hamdan Abdellatef, Lina J. Karam
Summary: This paper proposes performing the learning and inference processes in the compressed domain to reduce computational complexity and improve speed of neural networks. Experimental results show that modified ResNet-50 in the compressed domain is 70% faster than traditional spatial-based ResNet-50 while maintaining similar accuracy. Additionally, a preprocessing step with partial encoding is suggested to improve resilience to distortions caused by low-quality encoded images. Training a network with highly compressed data can achieve good classification accuracy with significantly reduced storage requirements.
Article
Computer Science, Artificial Intelligence
Victor R. Barradas, Yasuharu Koike, Nicolas Schweighofer
Summary: Inverse models are essential for human motor learning as they map desired actions to motor commands. The shape of the error surface and the distribution of targets in a task play a crucial role in determining the speed of learning.
Article
Computer Science, Artificial Intelligence
Ting Zhou, Hanshu Yan, Jingfeng Zhang, Lei Liu, Bo Han
Summary: We propose a defense strategy that reduces the success rate of data poisoning attacks in downstream tasks by pre-training a robust foundation model.
Article
Computer Science, Artificial Intelligence
Hao Sun, Li Shen, Qihuang Zhong, Liang Ding, Shixiang Chen, Jingwei Sun, Jing Li, Guangzhong Sun, Dacheng Tao
Summary: In this paper, the convergence rate of AdaSAM in the stochastic non-convex setting is analyzed. Theoretical proof shows that AdaSAM has a linear speedup property and decouples the stochastic gradient steps with the adaptive learning rate and perturbed gradient. Experimental results demonstrate that AdaSAM outperforms other optimizers in terms of performance.
Article
Computer Science, Artificial Intelligence
Juntong Yun, Du Jiang, Li Huang, Bo Tao, Shangchun Liao, Ying Liu, Xin Liu, Gongfa Li, Disi Chen, Baojia Chen
Summary: In this study, a dual manipulator grasping detection model based on the Markov decision process is proposed. By parameterizing the grasping detection model of dual manipulators using a cross entropy convolutional neural network and a full convolutional neural network, stable grasping of complex multiple objects is achieved. Robot grasping experiments were conducted to verify the feasibility and superiority of this method.
Article
Computer Science, Artificial Intelligence
Miaohui Zhang, Kaifang Li, Jianxin Ma, Xile Wang
Summary: This paper proposes an unsupervised person re-identification (Re-ID) method that uses two asymmetric networks to generate pseudo-labels for each other by clustering and updates and optimizes the pseudo-labels through alternate training. It also designs similarity compensation and similarity suppression based on the camera ID of pedestrian images to optimize the similarity measure. Extensive experiments show that the proposed method achieves superior performance compared to state-of-the-art unsupervised person re-identification methods.
Article
Computer Science, Artificial Intelligence
Florian Bacho, Dominique Chu
Summary: This paper proposes a new approach called the Forward Direct Feedback Alignment algorithm for supervised learning in deep neural networks. By combining activity-perturbed forward gradients, direct feedback alignment, and momentum, this method achieves better performance and convergence speed compared to other local alternatives to backpropagation.
Article
Computer Science, Artificial Intelligence
Xiaojian Ding, Yi Li, Shilin Chen
Summary: This research paper addresses the limitations of recursive feature elimination (RFE) and its variants in high-dimensional feature selection tasks. The proposed algorithms, which introduce a novel feature ranking criterion and an optimal feature subset evaluation algorithm, outperform current state-of-the-art methods.
Article
Computer Science, Artificial Intelligence
Naoko Koide-Majima, Shinji Nishimoto, Kei Majima
Summary: Visual images observed by humans can be reconstructed from brain activity, and the visualization of arbitrary natural images from mental imagery has been achieved through an improved method. This study provides a unique tool for directly investigating the subjective contents of the brain.
Article
Computer Science, Artificial Intelligence
Huanjie Tao, Qianyue Duan
Summary: In this paper, a hierarchical attention network with progressive feature fusion is proposed for facial expression recognition (FER), addressing the challenges posed by pose variation, occlusions, and illumination variation. The model achieves enhanced performance by aggregating diverse features and progressively enhancing discriminative features.
Article
Computer Science, Artificial Intelligence
Zhenyi Wang, Pengfei Yang, Linwei Hu, Bowen Zhang, Chengmin Lin, Wenkai Lv, Quan Wang
Summary: In the face of the complex landscape of deep learning, we propose a novel subgraph-level performance prediction method called SLAPP, which combines graph and operator features through an innovative graph neural network called EAGAT, providing accurate performance predictions. In addition, we introduce a mixed loss design with dynamic weight adjustment to improve predictive accuracy.
Article
Computer Science, Artificial Intelligence
Yiyang Yin, Shuangling Luo, Jun Zhou, Liang Kang, Calvin Yu-Chian Chen
Summary: Medical image segmentation is crucial for modern healthcare systems, especially in reducing surgical risks and planning treatments. Transanal total mesorectal excision (TaTME) has become an important method for treating colon and rectum cancers. Real-time instance segmentation during TaTME surgeries can assist surgeons in minimizing risks. However, the dynamic variations in TaTME images pose challenges for accurate instance segmentation.
Article
Computer Science, Artificial Intelligence
Teng Cheng, Lei Sun, Junning Zhang, Jinling Wang, Zhanyang Wei
Summary: This study proposes a scheme that combines the start-stop point signal features for wideband multi-signal detection, called Fast Spectrum-Size Self-Training network (FSSNet). By utilizing start-stop points to build the signal model, this method successfully solves the difficulty of existing deep learning methods in detecting discontinuous signals and achieves satisfactory detection speed.
Article
Computer Science, Artificial Intelligence
Wenming Wu, Xiaoke Ma, Quan Wang, Maoguo Gong, Quanxue Gao
Summary: The layer-specific modules in multi-layer networks are critical for understanding the structure and function of the system. However, existing methods fail to accurately characterize and balance the connectivity and specificity of these modules. To address this issue, a joint learning graph clustering algorithm (DRDF) is proposed, which learns the deep representation and discriminative features of the multi-layer network, and balances the connectivity and specificity of the layer-specific modules through joint learning.
Article
Computer Science, Artificial Intelligence
Guanghui Yue, Guibin Zhuo, Weiqing Yan, Tianwei Zhou, Chang Tang, Peng Yang, Tianfu Wang
Summary: This paper proposes a novel boundary uncertainty aware network (BUNet) for precise and robust colorectal polyp segmentation. BUNet utilizes a pyramid vision transformer encoder to learn multi-scale features and incorporates a boundary exploration module (BEM) and a boundary uncertainty aware module (BUM) to handle boundary areas. Experimental results demonstrate that BUNet outperforms other methods in terms of performance and generalization ability.