Pub Date : 2024-02-13DOI: 10.1109/TETCI.2024.3360336
Xiaoyuan Deng;Jinhai Li;Yuhua Qian;Junmin Liu
Fuzzy granular concepts are fundamental units in developing computational intelligence approaches based on fuzzy concept-cognitive learning. However, existing models in this field merely focus on the information provided by fuzzy granular concepts induced by objects, ignoring that of those induced by attributes. Consequently, these models underutilize the information provided by fuzzy granular concepts and weaken classification ability. To solve this problem, we propose an effective fuzzy granular concept-cognitive learning model, which incorporates fuzzy attribute granular concepts on the basis of the fuzzy object granular concepts. To be concrete, we firstly introduce the notion of a fuzzy attribute granular concept and construct a fuzzy granular concept space. Secondly, we obtain a fuzzy granular concept clustering space by optimizing the threshold which is used to fuse similar fuzzy granular concepts, and then form lower and upper approximation spaces through set approximation. In addition, we explain the mechanism of new incremental fuzzy concept-cognitive learning model for label prediction by integrating the fuzzy granular concept clustering space and the lower and upper approximation spaces. Finally, we show the classification performance of the proposed model on 28 datasets by comparing it with 10 classical machine learning classification algorithms and 17 fuzzy similarity-based classification algorithms, and evaluate incremental learning ability of our model. The experimental results demonstrate the feasibility and effectiveness of our method.
{"title":"An Emerging Incremental Fuzzy Concept-Cognitive Learning Model Based on Granular Computing and Conceptual Knowledge Clustering","authors":"Xiaoyuan Deng;Jinhai Li;Yuhua Qian;Junmin Liu","doi":"10.1109/TETCI.2024.3360336","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3360336","url":null,"abstract":"Fuzzy granular concepts are fundamental units in developing computational intelligence approaches based on fuzzy concept-cognitive learning. However, existing models in this field merely focus on the information provided by fuzzy granular concepts induced by objects, ignoring that of those induced by attributes. Consequently, these models underutilize the information provided by fuzzy granular concepts and weaken classification ability. To solve this problem, we propose an effective fuzzy granular concept-cognitive learning model, which incorporates fuzzy attribute granular concepts on the basis of the fuzzy object granular concepts. To be concrete, we firstly introduce the notion of a fuzzy attribute granular concept and construct a fuzzy granular concept space. Secondly, we obtain a fuzzy granular concept clustering space by optimizing the threshold which is used to fuse similar fuzzy granular concepts, and then form lower and upper approximation spaces through set approximation. In addition, we explain the mechanism of new incremental fuzzy concept-cognitive learning model for label prediction by integrating the fuzzy granular concept clustering space and the lower and upper approximation spaces. Finally, we show the classification performance of the proposed model on 28 datasets by comparing it with 10 classical machine learning classification algorithms and 17 fuzzy similarity-based classification algorithms, and evaluate incremental learning ability of our model. The experimental results demonstrate the feasibility and effectiveness of our method.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141096309","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-02-12DOI: 10.1109/TETCI.2024.3359042
Jianping Luo;Yongfei Dong;Qiqi Liu;Zexuan Zhu;Wenming Cao;Kay Chen Tan;Yaochu Jin
In this paper, we propose a multi-objective optimization algorithm based on multitask conditional neural processes (MTCNPs) to deal with expensive multi-objective optimization problems (MOPs). In the proposed algorithm, an MOP is decomposed into several subproblems. Several related subproblems are assigned to a task group and jointly handled using an MTCNPs surrogate model, in which multi-task learning is incorporated to exploit the similarity across the subproblems via joint surrogate model learning. Each subproblem in a task group is modeled by a conditional neural processes (CNPs) instead of a Gaussian Process (GP), thus avoiding the calculation of the GP covariance matrix. In addition, multiple subproblems are jointly learned through a multi-layer similarity network with activation function, which can measure and utilize the similarity and useful information among subproblems more effectively and improve the accuracy and robustness of the surrogate model. Experimental studies under several scenarios indicate that the proposed algorithm performs better than several state-of-the-art multi-objective evolutionary algorithms for expensive MOPs. The parameter sensitivity and effectiveness of the proposed algorithm are analyzed in detail.
本文提出了一种基于多任务条件神经过程(MTCNPs)的多目标优化算法,用于处理昂贵的多目标优化问题(MOPs)。在提议的算法中,一个 MOP 被分解成几个子问题。几个相关的子问题被分配到一个任务组,并使用 MTCNPs 代理模型进行联合处理,其中包含多任务学习,通过联合代理模型学习来利用各子问题之间的相似性。任务组中的每个子问题都采用条件神经过程(CNPs)建模,而不是高斯过程(GP),从而避免了计算 GP 协方差矩阵。此外,通过带激活函数的多层相似性网络对多个子问题进行联合学习,可以更有效地测量和利用子问题之间的相似性和有用信息,提高代用模型的准确性和鲁棒性。多种场景下的实验研究表明,针对昂贵的澳门威尼斯人官网程,所提出的算法比几种最先进的多目标进化算法性能更好。本文详细分析了所提算法的参数敏感性和有效性。
{"title":"A New Multitask Joint Learning Framework for Expensive Multi-Objective Optimization Problems","authors":"Jianping Luo;Yongfei Dong;Qiqi Liu;Zexuan Zhu;Wenming Cao;Kay Chen Tan;Yaochu Jin","doi":"10.1109/TETCI.2024.3359042","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3359042","url":null,"abstract":"In this paper, we propose a multi-objective optimization algorithm based on multitask conditional neural processes (MTCNPs) to deal with expensive multi-objective optimization problems (MOPs). In the proposed algorithm, an MOP is decomposed into several subproblems. Several related subproblems are assigned to a task group and jointly handled using an MTCNPs surrogate model, in which multi-task learning is incorporated to exploit the similarity across the subproblems via joint surrogate model learning. Each subproblem in a task group is modeled by a conditional neural processes (CNPs) instead of a Gaussian Process (GP), thus avoiding the calculation of the GP covariance matrix. In addition, multiple subproblems are jointly learned through a multi-layer similarity network with activation function, which can measure and utilize the similarity and useful information among subproblems more effectively and improve the accuracy and robustness of the surrogate model. Experimental studies under several scenarios indicate that the proposed algorithm performs better than several state-of-the-art multi-objective evolutionary algorithms for expensive MOPs. The parameter sensitivity and effectiveness of the proposed algorithm are analyzed in detail.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140309936","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-02-12DOI: 10.1109/TETCI.2024.3360303
Fengwei Gu;Jun Lu;Chengtao Cai;Qidan Zhu;Zhaojie Ju
Although Siamese trackers have become increasingly prevalent in the visual tracking domain, they are easily interfered by semantic distractors in complex environments, which results in the underutilization of feature information. Especially when multiple disturbances work together, the performance of many trackers often suffers severe degradation. To solve the above problem, this paper presents a robust Stereoscopic Transformer network for improving tracking performance. Using a hybrid attention mechanism, our method is composed of a channel feature awareness network (CFAN), a global channel attention network (GCAN), and a multi-level feature enhancement unit (MFEU). Concretely, CFAN focuses on specific channel information, while highlighting the contained target features and weakening the semantic distractor features. As an intermediate hub, GCAN is mainly responsible for establishing the global feature dependencies between the search region and the template, while selecting the concerned channel features to improve the distinguishing ability of the model. In particular, MFEU is used to enhance multi-level feature information to facilitate feature representation learning for our method. Finally, a Transformer-based Siamese tracker (named VTST) is proposed to present an efficient tracking representation, which can gain advantages over a variety of challenging attributes. Experiments show that our method outperforms the state-of-the-art trackers on multiple benchmarks with a real-time running speed of 56.0 fps.
{"title":"VTST: Efficient Visual Tracking With a Stereoscopic Transformer","authors":"Fengwei Gu;Jun Lu;Chengtao Cai;Qidan Zhu;Zhaojie Ju","doi":"10.1109/TETCI.2024.3360303","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3360303","url":null,"abstract":"Although Siamese trackers have become increasingly prevalent in the visual tracking domain, they are easily interfered by semantic distractors in complex environments, which results in the underutilization of feature information. Especially when multiple disturbances work together, the performance of many trackers often suffers severe degradation. To solve the above problem, this paper presents a robust Stereoscopic Transformer network for improving tracking performance. Using a hybrid attention mechanism, our method is composed of a channel feature awareness network (CFAN), a global channel attention network (GCAN), and a multi-level feature enhancement unit (MFEU). Concretely, CFAN focuses on specific channel information, while highlighting the contained target features and weakening the semantic distractor features. As an intermediate hub, GCAN is mainly responsible for establishing the global feature dependencies between the search region and the template, while selecting the concerned channel features to improve the distinguishing ability of the model. In particular, MFEU is used to enhance multi-level feature information to facilitate feature representation learning for our method. Finally, a Transformer-based Siamese tracker (named VTST) is proposed to present an efficient tracking representation, which can gain advantages over a variety of challenging attributes. Experiments show that our method outperforms the state-of-the-art trackers on multiple benchmarks with a real-time running speed of 56.0 fps.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141094880","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-02-12DOI: 10.1109/TETCI.2024.3359099
Shaik John Saida;Samit Ari
The accurate and consistent detection of ocean eddies significantly improves the monitoring of ocean surface dynamics and the identification of regional hydrographic and biological characteristics. The study of marine ecosystems and climate change requires an understanding of ocean eddies. Data from multi-satellite altimeters, which track sea surface height, are used in eddy detection. Altimeter measurements provide an accurate representation of the sea surface height. The existing deep learning-based eddy detection approaches suffer from high model and computational complexity. The fact that there are eddies of different diameters makes eddy identification more challenging. In this paper, the detection of ocean eddies using a dual encoder and decoder architecture is proposed to address these inadequacies. An attention mechanism is developed to comprehend the pixel-level context of the semantic segmentation. A series connection of separable convolutions is proposed to adequately describe the context of multi-scale fusion. Further, the tracking of eddies is also proposed using a novel tracking method. The experimental outcomes demonstrate that the proposed approach achieved mean intersection of union score, F-beta score, and mean pixel accuracy of 89.98 %, 94.47%, 95.13% and 89.66%, 94.54%, 95.51% on the Southern Atlantic Ocean and the South China Sea datasets.
{"title":"DUNet: Dual U-Net Architecture for Ocean Eddies Detection and Tracking","authors":"Shaik John Saida;Samit Ari","doi":"10.1109/TETCI.2024.3359099","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3359099","url":null,"abstract":"The accurate and consistent detection of ocean eddies significantly improves the monitoring of ocean surface dynamics and the identification of regional hydrographic and biological characteristics. The study of marine ecosystems and climate change requires an understanding of ocean eddies. Data from multi-satellite altimeters, which track sea surface height, are used in eddy detection. Altimeter measurements provide an accurate representation of the sea surface height. The existing deep learning-based eddy detection approaches suffer from high model and computational complexity. The fact that there are eddies of different diameters makes eddy identification more challenging. In this paper, the detection of ocean eddies using a dual encoder and decoder architecture is proposed to address these inadequacies. An attention mechanism is developed to comprehend the pixel-level context of the semantic segmentation. A series connection of separable convolutions is proposed to adequately describe the context of multi-scale fusion. Further, the tracking of eddies is also proposed using a novel tracking method. The experimental outcomes demonstrate that the proposed approach achieved mean intersection of union score, F-beta score, and mean pixel accuracy of 89.98 %, 94.47%, 95.13% and 89.66%, 94.54%, 95.51% on the Southern Atlantic Ocean and the South China Sea datasets.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140310223","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Reflection removal from an image with undesirable reflections is a challenging and ill-posed problem in low-level vision. In recent years, several deep learning approaches have been proposed to tackle the task of single image reflection removal (SIRR). These methods, however, do not fully utilize the fundamental image priors of reflection and lack interpretability. In this paper, we propose a deep variational inference reflection removal (VIRR) method for the SIRR problem, which has good interpretability and good generalization ability. Based on the proposed VIRR method, the posterior distributions of the latent transmission and reflection images can be estimated jointly through variational inference, using deep neural networks. Furthermore, the proposed network framework can be trained by the supervision of data-driven priors for the transmission image and reflection image, which is produced by the variational lower bound objective of marginal data likelihood. Our proposed method outperforms previous state-of-the-art approaches on four benchmark datasets, as demonstrated by extensive subjective and objective evaluations.
{"title":"Deep Variational Inference Network for Single Image Reflection Removal","authors":"Ya-Nan Zhang;Qiufu Li;Linlin Shen;Ailian He;Song Wu","doi":"10.1109/TETCI.2024.3359063","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3359063","url":null,"abstract":"Reflection removal from an image with undesirable reflections is a challenging and ill-posed problem in low-level vision. In recent years, several deep learning approaches have been proposed to tackle the task of single image reflection removal (SIRR). These methods, however, do not fully utilize the fundamental image priors of reflection and lack interpretability. In this paper, we propose a deep variational inference reflection removal (VIRR) method for the SIRR problem, which has good interpretability and good generalization ability. Based on the proposed VIRR method, the posterior distributions of the latent transmission and reflection images can be estimated jointly through variational inference, using deep neural networks. Furthermore, the proposed network framework can be trained by the supervision of data-driven priors for the transmission image and reflection image, which is produced by the variational lower bound objective of marginal data likelihood. Our proposed method outperforms previous state-of-the-art approaches on four benchmark datasets, as demonstrated by extensive subjective and objective evaluations.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140309998","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-02-12DOI: 10.1109/TETCI.2024.3359512
Jielong Chen;Yan Pan;Shuai Li;Yunong Zhang
Linear matrix-vector equations (LMVE) problem is widely encountered in science and engineering. Numerous methods have been proposed and studied to solve static (i.e., temporally-invariant) LMVE problem. However, many practical LMVE problems are temporally-variant. The static methods are not efficient and accurate enough. Originated from the research of Hopfield neuronet (HN), Zhang neuronet (ZN) is widely used to solve temporally-variant problems, but the traditional continuous ZN (TCZN) model needs to compute the inverse or pseudoinverse of the coefficient matrix, being less efficient. In this paper, a novel reciprocal ZN (RZN) model that does not need to compute the inverse or pseudoinverse of the coefficient matrix is proposed, and the detailed derivation procedure is first given. In addition, theoretical analyses show the global convergence performance of the RZN model. Moreover, the comparative numerical experiments with gradient neuronet (GN) model and TCZN model show the correctness and efficiency of RZN. Finally, the application of mobile localization further validates the superiority of RZN model over TCZN and GN models.
{"title":"Design and Analysis of Reciprocal Zhang Neuronet Handling Temporally-Variant Linear Matrix-Vector Equations Applied to Mobile Localization","authors":"Jielong Chen;Yan Pan;Shuai Li;Yunong Zhang","doi":"10.1109/TETCI.2024.3359512","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3359512","url":null,"abstract":"Linear matrix-vector equations (LMVE) problem is widely encountered in science and engineering. Numerous methods have been proposed and studied to solve static (i.e., temporally-invariant) LMVE problem. However, many practical LMVE problems are temporally-variant. The static methods are not efficient and accurate enough. Originated from the research of Hopfield neuronet (HN), Zhang neuronet (ZN) is widely used to solve temporally-variant problems, but the traditional continuous ZN (TCZN) model needs to compute the inverse or pseudoinverse of the coefficient matrix, being less efficient. In this paper, a novel reciprocal ZN (RZN) model that does not need to compute the inverse or pseudoinverse of the coefficient matrix is proposed, and the detailed derivation procedure is first given. In addition, theoretical analyses show the global convergence performance of the RZN model. Moreover, the comparative numerical experiments with gradient neuronet (GN) model and TCZN model show the correctness and efficiency of RZN. Finally, the application of mobile localization further validates the superiority of RZN model over TCZN and GN models.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140321689","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-02-12DOI: 10.1109/TETCI.2024.3359090
Jiahao Zheng;Yu Tang;Anthony Huang;Dapeng Wu
Face Sketch Recognition (FSR) is extremely challenging because of the heterogeneous gap between sketches and images. Relying on the ability to generative models, prior generation-based works have dominated FSR for a long time by decomposing FSR into two steps, namely, heterogeneous data synthesis and homogeneous data matching. However, decomposing FSR into two steps introduces noise and uncertainty, and the first step, heterogeneous data synthesis, is an even general and challenging problem. Solving a specific problem requires solving a more general one is to put the cart before the horse. In order to solve FSR smoothly and circumvent the above problems of generation-based methods, we propose a multi-view representation learning (MRL) framework based on Multivariate Loss and Hierarchical Loss (MvHi). Specifically, by using triplet loss as a bridge to connect the augmented representations generated by InfoNCE, we propose Multivariate Loss (Mv) to construct a more robust common feature subspace between sketches and images and directly solve FSR in this subspace. Moreover, Hierarchical Loss (Hi) is proposed to improve the training stability by utilizing the hidden states of the feature extractor. Comprehensive experiments on two commonly used datasets, CUFS and CUFSF, show that the proposed approach outperforms state-of-the-art methods by more than 7%. In addition, visualization experiments show that the proposed approach can extract the common representations among multi-view data compared to the baseline methods.
{"title":"Hierarchical Multivariate Representation Learning for Face Sketch Recognition","authors":"Jiahao Zheng;Yu Tang;Anthony Huang;Dapeng Wu","doi":"10.1109/TETCI.2024.3359090","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3359090","url":null,"abstract":"Face Sketch Recognition (FSR) is extremely challenging because of the heterogeneous gap between sketches and images. Relying on the ability to generative models, prior generation-based works have dominated FSR for a long time by decomposing FSR into two steps, namely, heterogeneous data synthesis and homogeneous data matching. However, decomposing FSR into two steps introduces noise and uncertainty, and the first step, heterogeneous data synthesis, is an even general and challenging problem. Solving a specific problem requires solving a more general one is to put the cart before the horse. In order to solve FSR smoothly and circumvent the above problems of generation-based methods, we propose a multi-view representation learning (MRL) framework based on Multivariate Loss and Hierarchical Loss (MvHi). Specifically, by using triplet loss as a bridge to connect the augmented representations generated by InfoNCE, we propose Multivariate Loss (Mv) to construct a more robust common feature subspace between sketches and images and directly solve FSR in this subspace. Moreover, Hierarchical Loss (Hi) is proposed to improve the training stability by utilizing the hidden states of the feature extractor. Comprehensive experiments on two commonly used datasets, CUFS and CUFSF, show that the proposed approach outperforms state-of-the-art methods by more than 7%. In addition, visualization experiments show that the proposed approach can extract the common representations among multi-view data compared to the baseline methods.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140321737","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-02-12DOI: 10.1109/TETCI.2024.3359096
Lei Li;Pan Liu;Chenyang Bu;Zan Zhang;Xindong Wu
As an emerging topic on preference learning, aiming at deducting the linear order of alternatives from the partial ranking, preference completion is to complete the preference of the target agent to form a linear order from the preferences of other agents under certain complex requirements. In order to improve the effectiveness and efficiency of preference completion in Big Data environments, firstly the preference graph is introduced to represent the collective preference of the agents over the alternatives with a certain consensus algorithm following the preference of the target agent. This preference graph can preserve rich information between agents. In addition, with the introduction of fuzzy ranking, it can illustrate the fuzziness of the target agent that can include several ranking options of the target agent over alternatives. Then, the satisfied preference can be matched from the preference graph with the fuzzy ranking requested by the target agent via isomorphism-based graph pattern matching. With the matched preference, the preference of the target agent can be completed. If the completed preference is not satisfied, the target agent can modify the fuzzy ranking, process the graph pattern rematching and complete the preference again. The experimental results show that with several real datasets the effectiveness and efficiency of the fuzzy ranking-based preference completion via graph pattern matching can be validated.
{"title":"Fuzzy Ranking-Based Preference Completion via Graph Pattern Matching and Rematching","authors":"Lei Li;Pan Liu;Chenyang Bu;Zan Zhang;Xindong Wu","doi":"10.1109/TETCI.2024.3359096","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3359096","url":null,"abstract":"As an emerging topic on preference learning, aiming at deducting the linear order of alternatives from the partial ranking, preference completion is to complete the preference of the target agent to form a linear order from the preferences of other agents under certain complex requirements. In order to improve the effectiveness and efficiency of preference completion in Big Data environments, firstly the preference graph is introduced to represent the collective preference of the agents over the alternatives with a certain consensus algorithm following the preference of the target agent. This preference graph can preserve rich information between agents. In addition, with the introduction of fuzzy ranking, it can illustrate the fuzziness of the target agent that can include several ranking options of the target agent over alternatives. Then, the satisfied preference can be matched from the preference graph with the fuzzy ranking requested by the target agent via isomorphism-based graph pattern matching. With the matched preference, the preference of the target agent can be completed. If the completed preference is not satisfied, the target agent can modify the fuzzy ranking, process the graph pattern rematching and complete the preference again. The experimental results show that with several real datasets the effectiveness and efficiency of the fuzzy ranking-based preference completion via graph pattern matching can be validated.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140310137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-02-12DOI: 10.1109/TETCI.2024.3359070
Hao Li;Dezhong Li;Maoguo Gong;Jianzhao Li;A. K. Qin;Lining Xing;Fei Xie
The traditional sparse unmixing methods based on multiobjective evolutionary algorithms (MOEAs) only deal with a single mixed pixel, without considering the spatial structure relationship between different mixed pixels. In addition, these methods suffer from the curse of dimensionality caused by the large number of pixels in hyperspectral image and spectra in library. In this paper, an evolutionary multitasking unmixing based on weakly nondominated sorting (EMTU-WNS) algorithm is proposed to alleviate these existing issues. First, a hyperspectral image is classified into multiple homogeneous regions, and the unmixing of pixels in the same region is constructed as a multiobjective optimization task. Then all the tasks are optimized simultaneously by using a population in the design of genetic transfer of intra-task and inter-task. In comparison with the original unmixing task with all pixels, these tasks in multiple homogeneous regions are relatively simple in term of dimensionality. Furthermore, it is inefficient for individuals to explore the whole search space. Therefore sparsity-constrained genetic operators are designed to evolve individuals towards the preference sparsity region. Finally, a preference-based weakly nondominated sorting is proposed to increase the number of nondominated solutions and maintain the diversity. The experimental results on three hyperspectral data sets demonstrate the effectiveness of EMTU-WNS with better convergence characteristics and unmixing accuracy.
{"title":"Sparse Hyperspectral Unmixing With Preference-Based Evolutionary Multiobjective Multitasking Optimization","authors":"Hao Li;Dezhong Li;Maoguo Gong;Jianzhao Li;A. K. Qin;Lining Xing;Fei Xie","doi":"10.1109/TETCI.2024.3359070","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3359070","url":null,"abstract":"The traditional sparse unmixing methods based on multiobjective evolutionary algorithms (MOEAs) only deal with a single mixed pixel, without considering the spatial structure relationship between different mixed pixels. In addition, these methods suffer from the curse of dimensionality caused by the large number of pixels in hyperspectral image and spectra in library. In this paper, an evolutionary multitasking unmixing based on weakly nondominated sorting (EMTU-WNS) algorithm is proposed to alleviate these existing issues. First, a hyperspectral image is classified into multiple homogeneous regions, and the unmixing of pixels in the same region is constructed as a multiobjective optimization task. Then all the tasks are optimized simultaneously by using a population in the design of genetic transfer of intra-task and inter-task. In comparison with the original unmixing task with all pixels, these tasks in multiple homogeneous regions are relatively simple in term of dimensionality. Furthermore, it is inefficient for individuals to explore the whole search space. Therefore sparsity-constrained genetic operators are designed to evolve individuals towards the preference sparsity region. Finally, a preference-based weakly nondominated sorting is proposed to increase the number of nondominated solutions and maintain the diversity. The experimental results on three hyperspectral data sets demonstrate the effectiveness of EMTU-WNS with better convergence characteristics and unmixing accuracy.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140310167","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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 : 2024-02-09DOI: 10.1109/TETCI.2024.3358981
Yujia Wang;Tong Wang;Chuang Li;Jiae Yang
In this study, we present a decentralized optimal fault-tolerant control (FTC) framework using neural networks (NNs) for interconnected nonlinear systems. This approach addresses challenges arising from unknown drift functions, interconnections, and multiple faults, including lock-in-place, loss of effectiveness, and float. Specifically, we propose a novel NN-based approximation scheme that utilizes a learning algorithm and a differentiator to estimate unknown information within the system. Additionally, our developed optimal control framework, in contrast to the conventional adaptive dynamic programming (ADP) approach, eliminates the need to separately design the optimal tracking controller into two parts, i.e., the steady-state controller and the feedback controller. Moreover, in the simulation section, control parameters are designed using the presented search algorithm, which demonstrates advantages in terms of both time efficiency and convenience. Finally, comparative simulations are conducted to illustrate the effectiveness of the proposed decentralized optimal fault-tolerant tracking control strategy.
在本研究中,我们针对互连非线性系统提出了一种使用神经网络(NN)的分散优化容错控制(FTC)框架。这种方法可以解决未知漂移函数、互连和多重故障(包括锁定、失效和浮动)带来的挑战。具体来说,我们提出了一种新颖的基于 NN 的近似方案,利用学习算法和微分器来估计系统内的未知信息。此外,与传统的自适应动态编程(ADP)方法相比,我们开发的最优控制框架无需将最优跟踪控制器分为稳态控制器和反馈控制器两部分进行设计。此外,在仿真部分,利用所介绍的搜索算法设计了控制参数,该算法在时间效率和便利性方面都具有优势。最后,通过对比仿真说明了所提出的分散式最优容错跟踪控制策略的有效性。
{"title":"Neural Network-Based Optimal Fault-Tolerant Control for Interconnected Nonlinear Systems With Actuator Failures","authors":"Yujia Wang;Tong Wang;Chuang Li;Jiae Yang","doi":"10.1109/TETCI.2024.3358981","DOIUrl":"https://doi.org/10.1109/TETCI.2024.3358981","url":null,"abstract":"In this study, we present a decentralized optimal fault-tolerant control (FTC) framework using neural networks (NNs) for interconnected nonlinear systems. This approach addresses challenges arising from unknown drift functions, interconnections, and multiple faults, including lock-in-place, loss of effectiveness, and float. Specifically, we propose a novel NN-based approximation scheme that utilizes a learning algorithm and a differentiator to estimate unknown information within the system. Additionally, our developed optimal control framework, in contrast to the conventional adaptive dynamic programming (ADP) approach, eliminates the need to separately design the optimal tracking controller into two parts, i.e., the steady-state controller and the feedback controller. Moreover, in the simulation section, control parameters are designed using the presented search algorithm, which demonstrates advantages in terms of both time efficiency and convenience. Finally, comparative simulations are conducted to illustrate the effectiveness of the proposed decentralized optimal fault-tolerant tracking control strategy.","PeriodicalId":13135,"journal":{"name":"IEEE Transactions on Emerging Topics in Computational Intelligence","volume":null,"pages":null},"PeriodicalIF":5.3,"publicationDate":"2024-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140310082","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}