This article investigates the adaptive neural network (NN) tracking control problem for nonlinear networked control systems (NCSs) with finite-time prescribed performance (FTPP) subject to intermittent denial-of-service (DoS) attacks. It is noticeable that when the DoS attacker is active, the controller does not receive any information, which makes the controller fail to work. To tackle the challenge, an adaptive NN switching state observer is first built to estimate the unmeasurable states. Second, an FTPP function is constructed to boost the transient and steady-state performances of NCSs. Third, under the framework of the backstepping technique, an adaptive command filter is established by combining the dynamic adaptive technique with the switching state observer, which handles the “complexity explosion” problem and improves the robustness of NCSs. Besides, it is rigorously proved mathematically that the boundedness of all signals in the closed-loop system and the designed controller compels the tracking error to fall into the predefined boundary within a finite time. Finally, an application-oriented example of the single-link robotic arm system is utilized to demonstrate the viability of the proposed control method.
{"title":"Observer-Based Adaptive NN Tracking Control for Nonlinear NCSs Under Intermittent DoS Attacks: A Finite-Time Prescribed Performance Method","authors":"Guangdeng Zong;Ruonan Liu;Hongzhen Xie;Yudi Wang;Xudong Zhao","doi":"10.1109/TSMC.2024.3521025","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3521025","url":null,"abstract":"This article investigates the adaptive neural network (NN) tracking control problem for nonlinear networked control systems (NCSs) with finite-time prescribed performance (FTPP) subject to intermittent denial-of-service (DoS) attacks. It is noticeable that when the DoS attacker is active, the controller does not receive any information, which makes the controller fail to work. To tackle the challenge, an adaptive NN switching state observer is first built to estimate the unmeasurable states. Second, an FTPP function is constructed to boost the transient and steady-state performances of NCSs. Third, under the framework of the backstepping technique, an adaptive command filter is established by combining the dynamic adaptive technique with the switching state observer, which handles the “complexity explosion” problem and improves the robustness of NCSs. Besides, it is rigorously proved mathematically that the boundedness of all signals in the closed-loop system and the designed controller compels the tracking error to fall into the predefined boundary within a finite time. Finally, an application-oriented example of the single-link robotic arm system is utilized to demonstrate the viability of the proposed control method.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2322-2331"},"PeriodicalIF":8.6,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143465796","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-01DOI: 10.1109/TSMC.2024.3520586
Dongdong Yue;Simone Baldi;Jinde Cao;Ling Shi
Due to the complex entanglement between distributed control and distributed estimation, adaptive multiagent dynamics over leaderless directed graphs are yet not completely understood. This happens even when the adaptive dynamics are based on established tools like model reference adaptive control (MRAC). This work starts from the observation that existing MRAC-based leaderless designs stop at asymptotic consensus, lacking of any guarantee for exponential consensus or parameter identification. The main contribution of this work is to show a new design departing from the existing ones in terms of exploiting persistence of excitation (PE): the proposed design is the first one attaining exponential consensus with exponential parameter identification over leaderless directed graphs. In the special case that the unknown parameters are homogeneous, PE can be relaxed to a weaker cooperative PE (C-PE) condition. The design is illustrated and verified alongside the state-of-the-art.
{"title":"Adaptive Exponential Consensus With Cooperative Exponential Parameter Identification Over Leaderless Directed Graphs","authors":"Dongdong Yue;Simone Baldi;Jinde Cao;Ling Shi","doi":"10.1109/TSMC.2024.3520586","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3520586","url":null,"abstract":"Due to the complex entanglement between distributed control and distributed estimation, adaptive multiagent dynamics over leaderless directed graphs are yet not completely understood. This happens even when the adaptive dynamics are based on established tools like model reference adaptive control (MRAC). This work starts from the observation that existing MRAC-based leaderless designs stop at asymptotic consensus, lacking of any guarantee for exponential consensus or parameter identification. The main contribution of this work is to show a new design departing from the existing ones in terms of exploiting persistence of excitation (PE): the proposed design is the first one attaining exponential consensus with exponential parameter identification over leaderless directed graphs. In the special case that the unknown parameters are homogeneous, PE can be relaxed to a weaker cooperative PE (C-PE) condition. The design is illustrated and verified alongside the state-of-the-art.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2354-2366"},"PeriodicalIF":8.6,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143465571","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-01DOI: 10.1109/TSMC.2024.3520322
Xuanxuan Ban;Jing Liang;Kunjie Yu;Yaonan Wang;Kangjia Qiao;Jinzhu Peng;Dunwei Gong;Canyun Dai
Evolutionary multitask optimization (EMTO) can solve multiple tasks simultaneously by leveraging the relevant information between tasks, but existing EMTO algorithms do not take into account the fact that almost all problems in the real world contain constraints. To address this dilemma, this article studies a local knowledge transfer-based evolutionary algorithm for constrained multitask optimization. To be specific, each task population is divided into multiple niches to enhance the diversity and control the intensity of knowledge transfer, thus avoiding excessive transfer of knowledge. Then a new similarity judgment method based on the information feedback of pioneer individuals is developed to judge the similarity between tasks and whether to perform knowledge transfer. Furthermore, two different transfer methods: a direct transfer and a learning transfer, are devised to perform knowledge transfer among niches pertaining to different tasks. In addition, an excellent-information-guided mutation mechanism is proposed to prevent niches from getting trapped in local optima and to promote rapid convergence. The system experiment on 18 constrained multitask test instances and 2 real-world problems demonstrate that the proposed algorithm outperforms or is at least comparable to other EMTO algorithms and constrained single-objective optimization algorithms.
{"title":"A Local Knowledge Transfer-Based Evolutionary Algorithm for Constrained Multitask Optimization","authors":"Xuanxuan Ban;Jing Liang;Kunjie Yu;Yaonan Wang;Kangjia Qiao;Jinzhu Peng;Dunwei Gong;Canyun Dai","doi":"10.1109/TSMC.2024.3520322","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3520322","url":null,"abstract":"Evolutionary multitask optimization (EMTO) can solve multiple tasks simultaneously by leveraging the relevant information between tasks, but existing EMTO algorithms do not take into account the fact that almost all problems in the real world contain constraints. To address this dilemma, this article studies a local knowledge transfer-based evolutionary algorithm for constrained multitask optimization. To be specific, each task population is divided into multiple niches to enhance the diversity and control the intensity of knowledge transfer, thus avoiding excessive transfer of knowledge. Then a new similarity judgment method based on the information feedback of pioneer individuals is developed to judge the similarity between tasks and whether to perform knowledge transfer. Furthermore, two different transfer methods: a direct transfer and a learning transfer, are devised to perform knowledge transfer among niches pertaining to different tasks. In addition, an excellent-information-guided mutation mechanism is proposed to prevent niches from getting trapped in local optima and to promote rapid convergence. The system experiment on 18 constrained multitask test instances and 2 real-world problems demonstrate that the proposed algorithm outperforms or is at least comparable to other EMTO algorithms and constrained single-objective optimization algorithms.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2183-2195"},"PeriodicalIF":8.6,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438447","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Trust relationships can facilitate cooperation in collective decisions. Using behavioral incentives via trust to encourage voluntary preference adjustments improves consensus through mutual agreement. This article aims to establish a trust incentive-driven framework for enabling consensus in social network group decision making (SN-GDM). First, a trust incentive mechanism is modeled via interactive trust functions that integrate risk attitude. The inclusion of risk attitude is crucial as it reflects the diverse ways decision makers (DMs) respond to uncertainty in trusting others’ judgments, capturing the varied behaviors of risky, neutral, and insurance DMs in the consensus process. Inconsistent DMs then adjust opinions in exchange for heightened trust. This mechanism enhances the importance degrees via a new weight assignment method, serving as a reward to motivate DMs to further align with the majority. Subsequently, a trust incentive-driven bounded maximum consensus model is proposed to optimize cooperation dynamics while preventing over-compensation of adjustments. Simulations and comparative analysis demonstrate the model’s efficacy in facilitating cooperation through tailored trust incentive mechanisms that account for these diverse risk preferences. Finally, the approach is applied to evaluate candidates for the Norden Shipping Scholarship, providing a cooperation-focused SN-GDM framework for achieving mutually agreeable solutions while acknowledging the impact of individual risk attitude on trust-based interactions.
{"title":"A Trust Incentive Driven Feedback Mechanism With Risk Attitude for Group Consensus in Social Networks","authors":"Feixia Ji;Jian Wu;Francisco Chiclana;Qi Sun;Enrique Herrera-Viedma","doi":"10.1109/TSMC.2024.3519510","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3519510","url":null,"abstract":"Trust relationships can facilitate cooperation in collective decisions. Using behavioral incentives via trust to encourage voluntary preference adjustments improves consensus through mutual agreement. This article aims to establish a trust incentive-driven framework for enabling consensus in social network group decision making (SN-GDM). First, a trust incentive mechanism is modeled via interactive trust functions that integrate risk attitude. The inclusion of risk attitude is crucial as it reflects the diverse ways decision makers (DMs) respond to uncertainty in trusting others’ judgments, capturing the varied behaviors of risky, neutral, and insurance DMs in the consensus process. Inconsistent DMs then adjust opinions in exchange for heightened trust. This mechanism enhances the importance degrees via a new weight assignment method, serving as a reward to motivate DMs to further align with the majority. Subsequently, a trust incentive-driven bounded maximum consensus model is proposed to optimize cooperation dynamics while preventing over-compensation of adjustments. Simulations and comparative analysis demonstrate the model’s efficacy in facilitating cooperation through tailored trust incentive mechanisms that account for these diverse risk preferences. Finally, the approach is applied to evaluate candidates for the Norden Shipping Scholarship, providing a cooperation-focused SN-GDM framework for achieving mutually agreeable solutions while acknowledging the impact of individual risk attitude on trust-based interactions.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2133-2146"},"PeriodicalIF":8.6,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438366","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this article, the twistors modeling method and high-order fully actuated (HOFA) system approach are utilized to investigate the attitude and orbit integrated control problem of spacecraft. Initially, a first-order state-space model is formulated to represent the attitude and orbit dynamics of spacecraft. This model is subsequently transformed into a HOFA system model. Based on this transformed model, a control strategy is meticulously designed. With the developed control strategy, a linear closed-loop system is obtained, whose poles can be arbitrarily configured. The effectiveness of the proposed control strategy is ultimately verified through detailed simulation results.
{"title":"Twistors-Based Attitude-Orbit Integrated Control for Spacecraft: A High-Order Fully Actuated System Approach","authors":"Dongyan Jin;Yannan Bi;Tong Wang;Jianbin Qiu;Huijun Gao","doi":"10.1109/TSMC.2024.3519436","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3519436","url":null,"abstract":"In this article, the twistors modeling method and high-order fully actuated (HOFA) system approach are utilized to investigate the attitude and orbit integrated control problem of spacecraft. Initially, a first-order state-space model is formulated to represent the attitude and orbit dynamics of spacecraft. This model is subsequently transformed into a HOFA system model. Based on this transformed model, a control strategy is meticulously designed. With the developed control strategy, a linear closed-loop system is obtained, whose poles can be arbitrarily configured. The effectiveness of the proposed control strategy is ultimately verified through detailed simulation results.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2098-2105"},"PeriodicalIF":8.6,"publicationDate":"2024-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438407","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article addresses the crucial aspect of safety in collaborative robotics by introducing a new continuous wavelet transform-convolutional neural network (CWT-CNN) for efficient robot collision detection. Unlike conventional methods, CWT-CNN exhibits superior data efficiency, requiring minimal collision data for robust training without relying on a dynamic model. The network’s adaptability extends to varying internal stiffness levels, offering robustness to changes in robotic system characteristics. Through comprehensive experimental studies, we investigate the impact of input signal types, wavelet types, wavelet scale ranges, and time-moving window sizes on collision detection performance, offering critical insights for optimal CWT parameter selection. Additionally, our transferability analysis demonstrates that the CWT-CNN can seamlessly adapt from one joint to another, requiring only minimal free-motion data from the new joint. This adaptability is validated through extensive experiments on an industrial robot and the robot equipped with variable stiffness actuators. In conclusion, the CWT-CNN is highly generalizable and data-efficient, making it a reliable solution for real-time collision detection in human-robot interactions, addressing a key aspect of safety in collaborative environments.
{"title":"Continuous Wavelet Network for Efficient and Transferable Collision Detection in Collaborative Robots","authors":"Zhenwei Niu;Taimur Hassan;Mohamed Nassim Boushaki;Naoufel Werghi;Irfan Hussain","doi":"10.1109/TSMC.2024.3518700","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3518700","url":null,"abstract":"This article addresses the crucial aspect of safety in collaborative robotics by introducing a new continuous wavelet transform-convolutional neural network (CWT-CNN) for efficient robot collision detection. Unlike conventional methods, CWT-CNN exhibits superior data efficiency, requiring minimal collision data for robust training without relying on a dynamic model. The network’s adaptability extends to varying internal stiffness levels, offering robustness to changes in robotic system characteristics. Through comprehensive experimental studies, we investigate the impact of input signal types, wavelet types, wavelet scale ranges, and time-moving window sizes on collision detection performance, offering critical insights for optimal CWT parameter selection. Additionally, our transferability analysis demonstrates that the CWT-CNN can seamlessly adapt from one joint to another, requiring only minimal free-motion data from the new joint. This adaptability is validated through extensive experiments on an industrial robot and the robot equipped with variable stiffness actuators. In conclusion, the CWT-CNN is highly generalizable and data-efficient, making it a reliable solution for real-time collision detection in human-robot interactions, addressing a key aspect of safety in collaborative environments.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2046-2061"},"PeriodicalIF":8.6,"publicationDate":"2024-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438362","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-31DOI: 10.1109/TSMC.2024.3519585
Yi Zhen;Xiao He;Donghua Zhou
This article proposes an intermittent optimal learning control strategy for nonlinear discrete-time systems under time-varying pass lengths and actuator faults. The target of the problem is to minimize the timewise tracking error and the input drifts, which are combined by a time-iteration-dependent factor. By searching the nearest available pass at each time instant for the current iteration, the optimal control gain can be obtained. Theoretical analysis indicates that the tracking error converges asymptotically in spite of the actuator fault and the robustness against the shifted initial state is further proven. Numerical simulations illustrate the effectiveness and robustness of the presented method.
{"title":"Optimal Learning Control for Nonlinear Faulty Systems With Time-Varying Trial Lengths","authors":"Yi Zhen;Xiao He;Donghua Zhou","doi":"10.1109/TSMC.2024.3519585","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3519585","url":null,"abstract":"This article proposes an intermittent optimal learning control strategy for nonlinear discrete-time systems under time-varying pass lengths and actuator faults. The target of the problem is to minimize the timewise tracking error and the input drifts, which are combined by a time-iteration-dependent factor. By searching the nearest available pass at each time instant for the current iteration, the optimal control gain can be obtained. Theoretical analysis indicates that the tracking error converges asymptotically in spite of the actuator fault and the robustness against the shifted initial state is further proven. Numerical simulations illustrate the effectiveness and robustness of the presented method.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2224-2236"},"PeriodicalIF":8.6,"publicationDate":"2024-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438487","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-31DOI: 10.1109/TSMC.2024.3519537
Zhenkun Wang;Lindong Xie;Genghui Li;Weifeng Gao;Maoguo Gong;Ling Wang
Surrogate-assisted evolutionary algorithms for addressing expensive optimization problems with both continuous and categorical variables (EOPCCVs) are still in the early stages of development. This study makes significant advancements by leveraging the mixed-variable nature of EOPCCVs in two crucial ways. First, it introduces a novel hybrid approach combining differential evolution and upper confidence bound sampling (DEUCB), designed to explore the mixed search space effectively. Second, a specialized value distance metric (VDM) is proposed, integrating continuous and categorical variables, to enhance the accuracy of the radial basis function (RBF) model approximation. Finally, we present a customized evolutionary expensive optimization algorithm (CEEO), which seamlessly incorporates DEUCB and RBF-VDM into the widely utilized global and local surrogate-assisted evolutionary optimization framework. Experimental results, compared against state-of-the-art counterparts on three distinct sets of benchmark problems and a convolutional neural network hyperparameter optimization task, consistently affirm the efficacy of the proposed CEEO in addressing EOPCCVs. The source code for the proposed CEEO algorithm is available at https://github.com/CIAM-Group/EvolutionaryAlgorithm_Codes/tree/main/CEEO_Code.
{"title":"Customized Evolutionary Expensive Optimization: Efficient Search and Surrogate Strategies for Continuous and Categorical Variables","authors":"Zhenkun Wang;Lindong Xie;Genghui Li;Weifeng Gao;Maoguo Gong;Ling Wang","doi":"10.1109/TSMC.2024.3519537","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3519537","url":null,"abstract":"Surrogate-assisted evolutionary algorithms for addressing expensive optimization problems with both continuous and categorical variables (EOPCCVs) are still in the early stages of development. This study makes significant advancements by leveraging the mixed-variable nature of EOPCCVs in two crucial ways. First, it introduces a novel hybrid approach combining differential evolution and upper confidence bound sampling (DEUCB), designed to explore the mixed search space effectively. Second, a specialized value distance metric (VDM) is proposed, integrating continuous and categorical variables, to enhance the accuracy of the radial basis function (RBF) model approximation. Finally, we present a customized evolutionary expensive optimization algorithm (CEEO), which seamlessly incorporates DEUCB and RBF-VDM into the widely utilized global and local surrogate-assisted evolutionary optimization framework. Experimental results, compared against state-of-the-art counterparts on three distinct sets of benchmark problems and a convolutional neural network hyperparameter optimization task, consistently affirm the efficacy of the proposed CEEO in addressing EOPCCVs. The source code for the proposed CEEO algorithm is available at <uri>https://github.com/CIAM-Group/EvolutionaryAlgorithm_Codes/tree/main/CEEO_Code</uri>.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2196-2210"},"PeriodicalIF":8.6,"publicationDate":"2024-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-30DOI: 10.1109/TSMC.2024.3519675
Yu Zhang;Xinyue Li;Wang Hu;Gary G. Yen
Symbolic regression is commonly considered in wide-ranging applications due to its inherent capability for learning both structure and weighting parameters of an interpretable model. However, for those scenarios that require fitting multiple expressions (MEs) synchronously, existing symbolic regression algorithms need to run multiple times step by step asynchronously for fitting such a group of expressions. Due to lacking mechanisms to explicitly capture and leverage the relationships between these expressions, the coupling information among MEs will be lost in this approach. A multiexpression symbolic regression algorithm (ME-SR) is proposed in this article to address the issue in learning MEs. Additionally, a methodology for extracting the approximate maximum common subexpression (aMCSE) among these MEs is suggested to disclose the relationships. A new metric is formulated to measure the quality of an aMCSE in ME-SR by imitating the concept of intersection over union. Furthermore, an adaptive cross matrix is incorporated into the algorithm to balance the search efforts between intertask and intratask domains. The proposed ME-SR demonstrates superior performance when compared to its counterparts of single expression symbolic regression on the designed test set. Finally, the efficacy of the method is well verified by a real-world circuit design case.
{"title":"Multiexpression Symbolic Regression and Its Circuit Design Case","authors":"Yu Zhang;Xinyue Li;Wang Hu;Gary G. Yen","doi":"10.1109/TSMC.2024.3519675","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3519675","url":null,"abstract":"Symbolic regression is commonly considered in wide-ranging applications due to its inherent capability for learning both structure and weighting parameters of an interpretable model. However, for those scenarios that require fitting multiple expressions (MEs) synchronously, existing symbolic regression algorithms need to run multiple times step by step asynchronously for fitting such a group of expressions. Due to lacking mechanisms to explicitly capture and leverage the relationships between these expressions, the coupling information among MEs will be lost in this approach. A multiexpression symbolic regression algorithm (ME-SR) is proposed in this article to address the issue in learning MEs. Additionally, a methodology for extracting the approximate maximum common subexpression (aMCSE) among these MEs is suggested to disclose the relationships. A new metric is formulated to measure the quality of an aMCSE in ME-SR by imitating the concept of intersection over union. Furthermore, an adaptive cross matrix is incorporated into the algorithm to balance the search efforts between intertask and intratask domains. The proposed ME-SR demonstrates superior performance when compared to its counterparts of single expression symbolic regression on the designed test set. Finally, the efficacy of the method is well verified by a real-world circuit design case.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2250-2263"},"PeriodicalIF":8.6,"publicationDate":"2024-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438364","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-27DOI: 10.1109/TSMC.2024.3517732
Wujie Zhou;Tingting Gong;Weiqing Yan
Deep-learning-based semantic segmentation has received increasing research attention in recent years. However, owing to complex architectures, existing approaches have failed to achieve high accuracies in real-time applications. In this article, a novel knowledge distillation (KD) SegFormer-based network, called KDSNet-S*, is proposed to explore the tradeoff between accuracy and efficiency. Specifically, a structured KD scheme is designed to transfer the rich advanced features of a teacher network (KDSNet-T) to a student network (KDSNet-S). Thereafter, the KDSNet-S network learns the precise segmentation ability of the KDSNet-T network. Additionally, a multifield perceptual fusion model is proposed to learn more integrated features for a single modality and obtain discriminative and comprehensive feature representations. Furthermore, a high-level feature integration module is introduced to refine multimodality high-level features. Finally, multilevel features are fused, and a label-decoupling-based three-stream decoder that decomposes the original semantic segmentation map into center and contour diffusion maps for different supervision tasks is introduced. Experimental results on two public red-green–blue-thermal semantic segmentation datasets indicate the superiority of KDSNet-S* over compared state-of-the-art methods. The KDSNet-S* reduces parameters and floating-point operations per second by 91.1% and 81.9%, respectively, compared with the KDSNet-T. The source codes and results will be available at https://github.com/purple-ting/KDSNet.
{"title":"Knowledge Distillation SegFormer-Based Network for RGB-T Semantic Segmentation","authors":"Wujie Zhou;Tingting Gong;Weiqing Yan","doi":"10.1109/TSMC.2024.3517732","DOIUrl":"https://doi.org/10.1109/TSMC.2024.3517732","url":null,"abstract":"Deep-learning-based semantic segmentation has received increasing research attention in recent years. However, owing to complex architectures, existing approaches have failed to achieve high accuracies in real-time applications. In this article, a novel knowledge distillation (KD) SegFormer-based network, called KDSNet-S*, is proposed to explore the tradeoff between accuracy and efficiency. Specifically, a structured KD scheme is designed to transfer the rich advanced features of a teacher network (KDSNet-T) to a student network (KDSNet-S). Thereafter, the KDSNet-S network learns the precise segmentation ability of the KDSNet-T network. Additionally, a multifield perceptual fusion model is proposed to learn more integrated features for a single modality and obtain discriminative and comprehensive feature representations. Furthermore, a high-level feature integration module is introduced to refine multimodality high-level features. Finally, multilevel features are fused, and a label-decoupling-based three-stream decoder that decomposes the original semantic segmentation map into center and contour diffusion maps for different supervision tasks is introduced. Experimental results on two public red-green–blue-thermal semantic segmentation datasets indicate the superiority of KDSNet-S* over compared state-of-the-art methods. The KDSNet-S* reduces parameters and floating-point operations per second by 91.1% and 81.9%, respectively, compared with the KDSNet-T. The source codes and results will be available at <uri>https://github.com/purple-ting/KDSNet</uri>.","PeriodicalId":48915,"journal":{"name":"IEEE Transactions on Systems Man Cybernetics-Systems","volume":"55 3","pages":"2170-2182"},"PeriodicalIF":8.6,"publicationDate":"2024-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438406","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：