Article
Computer Science, Information Systems
Chunliang Zhao, Yuren Zhou, Zefeng Chen
Summary: An automatic estimation mechanism based on the modified Ant Colony Algorithm is proposed in this study to assist the co evolution between subproblems, with working-ants executing local exploitation by recording subproblem information and commandants controlling global exploration by adjusting co-evolution. Experimental results demonstrate that the proposed algorithms outperform all referenced algorithms on multiple test suites.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Information Systems
Zhixia Zhang, Mengkai Zhao, Hui Wang, Zhihua Cui, Wensheng Zhang
Summary: This paper explores task scheduling in cloud computing and presents an interval many-objective optimization model and evolutionary algorithm, which consider uncertain factors while improving scheduling efficiency and performance.
INFORMATION SCIENCES
(2022)
Article
Automation & Control Systems
Huangke Chen, Ran Cheng, Witold Pedrycz, Yaochu Jin
Summary: This paper proposes a method to solve multiobjective optimization problems through multi-stage evolutionary search, highlighting convergence and diversity in different search stages. The algorithm balances and addresses the issues in multiobjective optimization through two stages.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2021)
Article
Computer Science, Information Systems
Maoqing Zhang, Lei Wang, Weian Guo, Wuzhao Li, Dongyang Li, Bo Hu, Qidi Wu
Summary: This paper proposes a relative non-dominance matrix and fitness formula to address the issue of dominance resistance in multi-objective optimization. Empirical analyses show that solutions with smaller fitness values are more likely to dominate other solutions in the evolutionary process and play a critical role in converging towards the true Pareto fronts. Additionally, the combination of k-means clustering strategy and the relative non-dominance matrix ensures diversity and adaptively adjusts the parameter k for environmental selection design.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Information Systems
Zhixia Zhang, Hui Wang, Wensheng Zhang, Zhihua Cui
Summary: A cooperative-competitive two-stage game mechanism assisted many-objective evolutionary algorithm (MaOEA-GM) is proposed to address the conflicts between convergence and diversity and the lack of Pareto selection pressure in many-objective optimization problems (MaOPs). The algorithm includes a competition stage with a strategy pool and a new game utility function to balance convergence and diversity, and a cooperative stage where individuals choose their preferred environmental selection mechanism through voting. Experimental results show that the MaOEA-GM algorithm outperforms five advanced MaOEAs in terms of convergence, diversity, and competitiveness in solving MaOPs.
INFORMATION SCIENCES
(2023)
Article
Computer Science, Artificial Intelligence
Jonathan E. Fieldsend, Tinkle Chugh, Richard Allmendinger, Kaisa Miettinen
Summary: Visualizing the search behavior of a series of points or populations in their native domain is crucial for understanding biases and attractors in an optimization process. This study introduces a distance-based many-objective optimization test problem that allows for visualization of search behavior in a 2-D design space. The authors' previous work further advances this research by providing a problem generator that can automatically create user-defined problem instances with various problem features.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Article
Computer Science, Artificial Intelligence
Xiangjuan Yao, Qian Zhao, Dunwei Gong, Song Zhu
Summary: This article proposes a method to solve large-scale many-objective optimization problems (LSMaOPs) based on dimension reduction and a solving knowledge-guided evolutionary algorithm (KGEA). The method effectively reduces the dimension of the original problem by clustering and aggregating the objective functions, and then solves the reduced problem using the solving KGEA. Experimental results demonstrate the effectiveness of the proposed algorithm in tackling LSMaOPs.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2023)
Article
Computer Science, Artificial Intelligence
Sanyan Chen, Xuewu Wang, Jin Gao, Wei Du, Xingsheng Gu
Summary: The paper introduces an adaptive switching strategy-based evolutionary algorithm to address the selection pressure and diversity issues in many-objective optimization. The algorithm dynamically switches between two deletion criteria in each generation to effectively remove poor solutions, demonstrating its effectiveness and advantages through comparisons with state-of-the-art algorithms on benchmark problems.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Automation & Control Systems
Songbai Liu, Qiuzhen Lin, Kay Chen Tan, Maoguo Gong, Carlos A. Coello Coello
Summary: This article proposes a fuzzy decomposition-based MOEA that estimates the population's shape using fuzzy prediction and selects weight vectors to fit the Pareto front shapes of different multi-objective optimization problems.
IEEE TRANSACTIONS ON CYBERNETICS
(2022)
Article
Computer Science, Artificial Intelligence
Jiawei Yuan, Hai-Lin Liu, Fangqing Gu, Qingfu Zhang, Zhaoshui He
Summary: This article investigates the properties of ratio and difference-based indicators under the Minkovsky distance, and proposes an algorithm for solution evaluation using a ratio-based indicator. By identifying promising regions and ensuring population diversity, the algorithm demonstrates competitive performance on various benchmark problems.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2021)
Article
Computer Science, Information Systems
Songbai Liu, Junhao Zheng, Qiuzhen Lin, Kay Chen Tan
Summary: This paper proposes an MOEA using clustering with a flexible similarity metric to tackle multi and many-objective optimization problems with irregular Pareto fronts. By predicting the concavity or convexity of the problem and setting a flexible reference point, the algorithm can properly measure the similarity between solutions and classify them effectively in environmental selection, showing significant advantages in experimental results.
INFORMATION SCIENCES
(2021)
Article
Mathematics
Chengxin Wen, Hongbin Ma
Summary: Many-objective optimization is an important research topic in evolutionary computing, and a two-stage hypervolume-based evolutionary algorithm is proposed to achieve convergence and diversity through global and local searches. Experimental results show that the algorithm is competitive in most cases.
Article
Mathematics
Yizhang Xia, Jianzun Huang, Xijun Li, Yuan Liu, Jinhua Zheng, Juan Zou
Summary: This paper discusses the balance between convergence and diversity in many-objective evolutionary optimization algorithms. The authors propose a new algorithm called Indicator and Decomposition-based Evolutionary Algorithm (IDEA) to achieve both convergence and diversity. Experimental results show that IDEA outperforms other state-of-the-art many-objective algorithms.
Article
Computer Science, Artificial Intelligence
Zhe Liu, Fei Han, Qinghua Ling, Henry Han, Jing Jiang
Summary: This paper proposes a many-objective optimization evolutionary algorithm based on the hyper-dominance degree. It quantifies the convergence of each solution using hyper-dominance degree and balances convergence and diversity through tolerance adjusting, reference vectors-based diversity preservation, and population reselection strategies.
SWARM AND EVOLUTIONARY COMPUTATION
(2023)
Article
Engineering, Chemical
Funan Peng, Li Lv, Weiru Chen, Jun Wang
Summary: In this paper, a projection-based evolutionary algorithm called MOEA/PII is introduced, which divides the objective space into projection plane and free dimension(s) using the idea of dimension reduction and decomposition. The balance between convergence and diversity is maintained using a Bi-Elite queue. MOEA/PII is an algorithm framework that can be combined with other decomposition-based or dominance-based algorithms, showing better performance.
Article
Computer Science, Artificial Intelligence
Zhengqiang Zhang, Qinmu Peng, Sichao Fu, Wenjie Wang, Yiu-Ming Cheung, Yue Zhao, Shujian Yu, Xinge You
Summary: In this article, the author proposes an improved method for weakly supervised semantic segmentation. By decomposing the position information into high-level semantic information and low-level physical information, the author develops corresponding mechanisms to recover each component independently. Experimental results show that the proposed dual-feedback network outperforms existing state-of-the-art methods in terms of mIoU.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Cuie Yang, Yiu-Ming Cheung, Jinliang Ding, Kay Chen Tan, Bing Xue, Mengjie Zhang
Summary: This work addresses the problem of unsupervised partial domain adaptation (PDA) and proposes a contrastive learning-assisted alignment (CLA) approach to enhance adaptation and reduce the contribution of outlier instances. It utilizes contrastive losses for cluster matching and introduces a new reweighting scheme for weight estimation.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Yuqun Yang, Xu Tang, Yiu-Ming Cheung, Xiangrong Zhang, Licheng Jiao
Summary: The classification of high-resolution remote sensing (HRRS) images is a challenging task due to the diverse and massive content within these images. Existing models for HRRS scene classification often treat it as a single-label problem, ignoring the various semantics hidden in the images and leading to inaccurate decisions. To address this, we propose a semantic-aware graph network (SAGN) that utilizes dense feature pyramid networks, adaptive semantic analysis modules, dynamic graph feature update modules, and scene decision modules to extract multi-scale information, mine various semantics, and make accurate decisions. Our extensive experiments on popular HRRS scene datasets demonstrate the effectiveness of SAGN.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2023)
Article
Computer Science, Artificial Intelligence
Fangqing Gu, Haosen Liu, Yiu-ming Cheung, Hai -Lin Liu
Summary: This study proposes an adaptive constraint regulation method to balance the feasibility and convergence of solutions by adjusting the constraint violation of infeasible solutions. Experimental results demonstrate that the proposed method effectively achieves solution balance and improves solution diversity.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Meng Pang, Binghui Wang, Mang Ye, Yiu-ming Cheung, Yiran Chen, Bihan Wen
Summary: Single sample per person face recognition is a challenging problem, and the existing methods have limitations. Therefore, a novel model was proposed to overcome these limitations, and experiments show its superiority.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Yichao Tang, Shuai Wang, Chuntao Wang, Shijun Xiang, Yiu-Ming Cheung
Summary: In this paper, a two-stage robust reversible watermarking (RRW) scheme is proposed to improve robustness and capacity. The first stage inserts a robust watermark into selected Pseudo-Zernike moments (PZMs) using an adaptive normalization method and an optimized embedding strategy. The second stage embeds a reversible watermark to achieve reversibility.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Automation & Control Systems
Hao-Tian Wu, Yiu-Ming Cheung, Zhenwei Zhuang, Lingling Xu, Jiankun Hu
Summary: Reversible data hiding in ciphertext has potential applications for privacy protection and transmitting extra data in a cloud environment. However, applying homomorphic processing to an encrypted image with hidden data is challenging due to possible changes in image content caused by preprocessing or/and data embedding. To address this issue, a lossless data hiding method called random element substitution (RES) is proposed, which replaces the to-be-hidden bits with the random element of a cipher value. The RES method is combined with another preprocessing-free algorithm to generate two schemes for lossless data hiding in encrypted images.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Computer Science, Artificial Intelligence
Mengke Li, Yiu-ming Cheung
Summary: This paper proposes an identity-preserved complete face recovering (CFR) approach to recover the full face image of a target person from a partial face image. The approach utilizes a denoising auto-encoder based network and an adversarial structure with a new variant discriminator. Experimental results show the superiority of the proposed method.
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Wei Li, Bo Sun, Yafeng Sun, Ying Huang, Yiu-ming Cheung, Fangqing Gu
Summary: In this paper, a diversity controller (DC) based on a small-world network and an infeasible-feasible regions constraint handling method (IF) are proposed. They are applied to the success-history-based parameter adaptive differential evolution (SHADE) algorithm, resulting in a DC-SHADE-IF algorithm. Experimental results demonstrate the superior performance of the proposed algorithm in terms of accuracy and convergence.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Yiqun Zhang, Yiu-Ming Cheung
Summary: This article addresses the challenge of defining similarity between data objects with heterogeneous attributes. It proposes a new dissimilarity metric that computes the dissimilarities between attribute values using graph structures. The article also introduces a new k-means-type clustering algorithm associated with this metric, which is capable of analyzing datasets composed of various combinations of attribute types.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Hongmin Cai, Xiaoqi Sheng, Guorong Wu, Bin Hu, Yiu-Ming Cheung, Jiazhou Chen
Summary: There is increasing evidence that Alzheimer's disease (AD) disrupts the brain network before clinical symptoms appear, allowing for early diagnosis. The current methods of analyzing brain networks treat the high-dimensional data as regular matrices or vectors, which leads to a loss of essential network topology and affects diagnosis accuracy. To address this issue, this article proposes a network manifold harmonic discriminant analysis (MHDA) method for accurately detecting AD. The effectiveness of the proposed method in stratifying cognitively normal controls, mild cognitive impairment, and AD is demonstrated through extensive experiments.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Chao Yang, Qiang Liu, Yi Liu, Yiu-Ming Cheung
Summary: This article proposes a novel dynamic latent variable (DLV)-based transfer learning approach, called transfer DLV regression (TDLVR), for quality prediction of multimode processes with dynamics. It can extract the dynamics between process variables and quality variables in the principal operating mode (POM) and also the co-dynamic variations among process variables between the POM and the new mode. An error compensation mechanism is incorporated to adapt to the conditional distribution discrepancy and make full use of the available labeled samples from the new mode.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Shu-Juan Peng, Ye Fan, Yiu-ming Cheung, Xin Liu, Zhen Cui, Taihao Li
Summary: This paper proposes an efficient deep cross-modal anomaly detection approach via TN-BCL, which can effectively identify various cross-modal anomalies within heterogeneous multi-modal data.
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Mengke Li, Yiu-Ming Cheung, Zhikai Hu
Summary: For long-tailed distributed data, existing classification models often focus on the head classes and neglect the tail classes, resulting in poor generalization performance. To address this issue, a new approach is proposed in this paper, which introduces a key point sensitive (KPS) loss to enhance the generalization capability of the classification model by regularizing the key points. Additionally, the proposed approach assigns larger margins on the tail classes to improve their performance.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Wei Huang, Yintao Zhou, Yiu-ming Cheung, Peng Zhang, Yufei Zha, Meng Pang
Summary: Parkinson's disease is a neurodegenerative disease that can be diagnosed through facial expressions. However, existing methods are limited by training data and prediction model performance. To address these limitations, we propose a facial expression-guided PD diagnosis method based on high-quality training data augmentation and deep neural network prediction.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)