Article
Computer Science, Artificial Intelligence
Hao Xu, A. K. Qin, Siyu Xia
Summary: Evolutionary Multitask Optimization (EMTO) uses evolutionary algorithms (EAs) to solve multiple optimization tasks simultaneously, utilizing knowledge transfer to improve performance. The proposed adaptive EMTO (AEMTO) framework adjusts knowledge transfer in a synergistic way, effectively addressing negative knowledge transfer and enhancing overall performance.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Article
Computer Science, Artificial Intelligence
Chao Lyu, Yuhui Shi, Lijun Sun, Chin-Teng Lin
Summary: This article proposes a novel algorithm for community detection in multiplex networks. The algorithm decomposes the problem into two parts, detecting specific community partitions for each component layer and finding the composite community structure shared by all layers. Experimental results demonstrate that the algorithm outperforms classical and state-of-the-art algorithms in community detection on multiplex networks.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2023)
Article
Computer Science, Artificial Intelligence
Zedong Tang, Maoguo Gong, Yue Wu, Wenfeng Liu, Yu Xie
Summary: The article presents a novel and computationally efficient intertask information transfer strategy by aligning subspaces. By introducing a learnable alignment matrix, it extracts complementary information among different tasks to enhance the performance of solving complicated problems. This method shows superior performance compared to existing evolutionary multitask optimization algorithms in comprehensive experiments.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2021)
Review
Computer Science, Artificial Intelligence
Tingyang Wei, Shibin Wang, Jinghui Zhong, Dong Liu, Jun Zhang
Summary: This paper presents a detailed exposition on the research in the field of evolutionary multitask optimization (EMTO), revealing the core components of EMTO algorithms and the fusion between EMTO and traditional evolutionary algorithms. By analyzing the associations of different strategies in various branches of EMTO, this review uncovers research trends and potentially important directions, as well as mentions interesting real-world applications.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Article
Automation & Control Systems
Yi Jiang, Zhi-Hui Zhan, Kay Chen Tan, Jun Zhang
Summary: This article proposes a block-level knowledge transfer (BLKT) framework to overcome the limitations of knowledge transfer in multitask optimization problems. BLKT divides individuals into blocks and transfers knowledge at the block-level, enabling transfer between similar dimensions belonging to the same or different tasks. Extensive experiments show that BLKT-based differential evolution outperforms state-of-the-art algorithms in multitask optimization and also achieves competitive performance in single-task global optimization.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Computer Science, Artificial Intelligence
Jian-Yu Li, Zhi-Hui Zhan, Kay Chen Tan, Jun Zhang
Summary: Knowledge transfer plays a vital role in solving multitask optimization problems. This article proposes a meta-knowledge transfer-based differential evolution (MKTDE) algorithm, which efficiently solves MTOPs using a more general approach. By transferring meta-knowledge, the MKTDE algorithm effectively associates different tasks' heterogeneous multisource data to solve MTOPs more efficiently. Two novel methods, multiple populations for the multiple tasks framework and elite solution transfer, further enhance the MKTDE algorithm. Extensive experiments validate the superior performance of the proposed algorithm compared to state-of-the-art approaches.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Article
Computer Science, Artificial Intelligence
Kangjia Qiao, Kunjie Yu, Boyang Qu, Jing Liang, Hui Song, Caitong Yue
Summary: This article presents an evolutionary multitasking-based constrained multiobjective optimization framework for solving CMOPs. It transforms the optimization problem into two related tasks and utilizes a tentative method to discover and transfer useful knowledge. The approach achieves better performance compared to other state-of-the-art algorithms.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Article
Computer Science, Artificial Intelligence
Xianpeng Wang, Zhiming Dong, Lixin Tang, Qingfu Zhang
Summary: This article proposes a multiobjective multitask optimization algorithm based on decomposition with dual neighborhoods, which improves algorithm performance by transferring knowledge among different tasks through neighborhood usage.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2023)
Article
Computer Science, Artificial Intelligence
Honggui Han, Xing Bai, Huayun Han, Ying Hou, Junfei Qiao
Summary: This article proposes a self-adjusting multitask particle swarm optimization algorithm to address the problem of negative transfer in multitask optimization. By designing an effective knowledge estimation metric and a self-adjusting knowledge transfer mechanism, the algorithm achieves effective knowledge transfer and removes ineffective knowledge. Convergence analysis is provided to guarantee the effectiveness of the algorithm. Experimental results demonstrate that the proposed algorithm outperforms other algorithms in suppressing negative transfer and achieving convergence.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Article
Computer Science, Artificial Intelligence
Sheng-Hao Wu, Zhi-Hui Zhan, Kay Chen Tan, Jun Zhang
Summary: This article proposes a novel orthogonal transfer (OT) method enabled by a cross-task mapping (CTM) strategy, which achieves high-quality knowledge transfer among heterogeneous tasks. The OT method handles task dimensionality differences and finds the best combination of different dimensions for high-quality knowledge transfer.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2023)
Article
Computer Science, Artificial Intelligence
Kangjia Qiao, Jing Liang, Kunjie Yu, Minghui Wang, Boyang Qu, Caitong Yue, Yinan Guo
Summary: This paper proposes a double-balanced evolutionary multi-task optimization (DBEMTO) algorithm to better solve constrained multi-objective optimization problems (CMOPs). DBEMTO evolves two populations to solve the main task (CMOP) and the auxiliary task (MOP extracted from the CMOP) respectively and uses three evolutionary strategies for offspring generation. DBEMTO has performed more competitively compared to other state-of-the-art CMOEAs according to the final results.
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Aritz D. Martinez, Javier Del Ser, Eneko Osaba, Francisco Herrera
Summary: This paper introduces an adaptive multitask reinforcement learning algorithm called A-MFEA-RL, which improves performance by facilitating the exchange of genetic material through crossover and inheritance mechanisms. Experimental results show that A-MFEA-RL achieves high success rates when handling multiple tasks and enhances knowledge exchange among tasks.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Article
Automation & Control Systems
Zhengping Liang, Weiqi Liang, Zhiqiang Wang, Xiaoliang Ma, Ling Liu, Zexuan Zhu
Summary: The EMT-PD algorithm improves convergence performance by adjusting search step size and dynamically changing search range based on population distribution. This two-stage adaptive knowledge transfer approach reduces negative transfer effects and enhances population diversity, helping to escape local optima. Experimental results demonstrate the superiority of EMT-PD in multitasking multiobjective optimization.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Thiago Rios, Bas van Stein, Thomas Back, Bernhard Sendhoff, Stefan Menzel
Summary: The choice of design representations and search operators is crucial for the performance of multitask optimization algorithms. Mapping the design space of each task to a common search space is challenging in engineering cases. This study applies a 3D point cloud autoencoder to map design representations to the latent space, improving the performance of multitask optimization.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Review
Computer Science, Artificial Intelligence
Ziying Tan, Linbo Luo, Jinghui Zhong
Summary: Evolutionary multi-task optimization (EMTO) is an optimization algorithm that aims to optimize multiple tasks simultaneously. It utilizes common knowledge across tasks to improve performance in solving each task independently. This survey focuses on the research progress of knowledge transfer methods in EMTO and proposes a taxonomy to categorize the existing work. It aims to identify research directions for improving knowledge transfer performance in EMTO.
APPLIED SOFT COMPUTING
(2023)
Article
Automation & Control Systems
Chaolong Ying, Jing Liu, Kai Wu, Chao Wang
Summary: This article proposes a subspace learning-based evolutionary multiobjective network reconstruction algorithm, which utilizes logistic principal component analysis (LPCA) to learn a subspace containing the features of the network structure and utilizes a preference-based strategy to concentrate on finding solutions approximate to the true sparsity. The experimental results demonstrate the effectiveness of the proposed method in reconstructing large-scale networks.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Computer Science, Artificial Intelligence
Wang Chao, Kai Wu, Jing Liu
Summary: Learning how to optimize AUC performance for imbalanced data has been a topic of interest. This paper proposes an evolutionary multitasking framework (EMTAUC) that utilizes information from different tasks to improve AUC performance.
IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE
(2022)
Article
Computer Science, Artificial Intelligence
Kai Wu, Kaixin Yuan, Yingzhi Teng, Jing Liu, Licheng Jiao
Summary: This paper proposes a time series classification method based on the broad fuzzy cognitive map system (BFCMS), which includes a feature extraction block, a spatiotemporal information aggregation block, and a prediction layer. BFCMS achieves efficient time series classification by addressing the limitations of broad learning systems. Experimental results demonstrate the superiority of BFCMS.
APPLIED SOFT COMPUTING
(2022)
Article
Automation & Control Systems
Kai Wu, Jing Liu
Summary: This paper investigates the problem of learning large-scale fuzzy cognitive maps with a limited computational budget. The authors propose two strategies to address this problem and demonstrate the effectiveness of the proposed methods through experiments.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2022)
Article
Computer Science, Artificial Intelligence
Chao Wang, Jiaxuan Zhao, Lingling Li, Licheng Jiao, Jing Liu, Kai Wu
Summary: Influence maximization is crucial for mining deep information in social networks by selecting a seed set to maximize the number of influenced nodes. Existing studies propose alternate transformations with lower computational costs to evaluate influence spread. This article presents a multi-transformation evolutionary framework for influence maximization (MTEFIM) to exploit the potential similarities and advantages of alternate transformations. MTEFIM optimizes multiple transformations simultaneously as tasks and achieves highly competitive performance compared to popular IM-specific methods.
IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE
(2023)
Article
Automation & Control Systems
Kai Wu, Chao Wang, Jing Liu
Summary: This article introduces an evolutionary multitasking multilayer network reconstruction framework EM2MNR to enhance reconstruction performance by utilizing correlations among different component layers. By utilizing restricted Boltzmann machine to extract low effective features and deciding on knowledge transfer, the proposed framework significantly improves reconstruction performance on multilayer network reconstruction problems.
IEEE TRANSACTIONS ON CYBERNETICS
(2022)
Review
Computer Science, Artificial Intelligence
Hong Zhao, Xuhui Ning, Xiaotao Liu, Chao Wang, Jing Liu
Summary: Evolutionary multi-task optimization (EMTO) is a new branch of evolutionary algorithm (EA) that aims to optimize multiple tasks simultaneously within a same problem and output the best solution for each task. EMTO utilizes the strengths of EA to perform global optimization without relying on the mathematical properties of the problem. Therefore, EMTO is particularly suitable for complex, non-convex and nonlinear problems.
APPLIED SOFT COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
Yitong Li, Kai Wu, Jing Liu
Summary: This paper proposes a robust time series prediction framework called spARIMA, which reduces noise interference by designing a sequential training scheme in batches based on the degree of noise. spARIMA relies on the differential prediction model in ARIMA and absorbs the advantages of the gradual training scheme in self-paced learning (SPL) to effectively address the instability caused by noise. Furthermore, spARIMA introduces diversity selection to avoid selecting similar samples, using a weighted local complexity-similarity distance expression to represent the diversity of noisy data. Comparative tests with existing ARIMA models on two gradient descent algorithms show that spARIMA not only works well with noisy data, but also performs efficiently with normal data, indicating its generalization ability.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Physics, Multidisciplinary
Yuanyuan Li, Kai Wu, Jing Liu
Summary: This paper discusses the method of discovering governing equations from streaming data. An online modeling method is proposed to handle samples one by one sequentially and performs well in discovering ordinary and partial differential equations. Additionally, the proposed method is competitive in identifying change points and discovering governing equations.
PHYSICAL REVIEW RESEARCH
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Kai Wu, Xiangyi Teng, Jing Liu
Summary: This article proposes a method using fuzzy cognitive map technique to identify the enemy's hidden military power, and verifies its effectiveness through experiments. The algorithm locates hidden nodes by measuring the anomalies between fuzzy cognitive maps obtained from different data segments.
COMPUTER SUPPORTED COOPERATIVE WORK AND SOCIAL COMPUTING, CHINESECSCW 2021, PT II
(2022)