Article
Computer Science, Artificial Intelligence
Jong Hyok Ri, Tok Gil Kang, Chol Ryong Choe
Summary: This paper focuses on a semi-supervised transfer learning algorithm called target adaptive ELM (TAELM) under the ELM framework. The proposed method learns a high-quality target-specific classifier by introducing a knowledge transfer term and a graph laplacian-based manifold regularization term. Experimental results show that the proposed approach significantly outperforms other state-of-the-art algorithms with fewer resources.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2023)
Article
Automation & Control Systems
Bernardo Fichera, Aude Billard
Summary: Dynamical Systems are fundamental for modeling and understanding time evolving phenomena, and data-driven approaches are preferred for identifying and controlling nonlinear DS. This paper proposes a graph-based spectral clustering method that takes advantage of a velocity-augmented kernel to connect data points belonging to the same dynamics, while preserving the natural temporal evolution.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Automation & Control Systems
Le Yang, Shiji Song, Shuang Li, Yiming Chen, Gao Huang
Summary: The proposed GDR-ELM, a graph embedding-based DR framework, reconstructs all samples according to the weights in a graph matrix containing supervised information, instead of self-reconstruction. GDR-ELM can be stacked as building blocks to construct a multilayer framework for more complicated representation learning tasks, and experiments on various datasets demonstrate its effectiveness.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Liang-Rui Ren, Jin-Xing Liu, Ying-Lian Gao, Xiang-Zhen Kong, Chun-Hou Zheng
Summary: KRSL is a nonlinear similarity measure defined in kernel space, which enhances the accuracy of gradient based methods while weakening the negative effects of noise and outliers. Incorporating KRSL into ELM can improve its ability to handle noisy and outlier-laden data and has shown promising results in robust machine learning and signal processing.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Engineering, Electrical & Electronic
Xiangyu Liu, Peng Song, Chao Sheng, Wenjing Zhang
Summary: Non-negative matrix factorization (NMF) has been widely studied for multi-view clustering. Existing methods often neglect the high-order relationship among features and the complementary information in different views. In this paper, we propose a robust multi-view NMF approach that considers both high-order relationship and complementary information, achieving effective clustering results.
DIGITAL SIGNAL PROCESSING
(2022)
Article
Computer Science, Artificial Intelligence
Zhiqiang Zeng, Xiaodong Wang, Wei Li, Yuandi Ye
Summary: In natural scene classification, multi-label learning is a popular research topic due to the fact that one image can belong to multiple categories simultaneously. Conventional multi-label learning models face challenges when dealing with high-dimensional data, such as redundant features and high computational costs. To address this, we propose an alternative multi-label feature learning solution that incorporates both labeled and unlabeled information. By extracting label correlation and propagating label information, we are able to discover noise or outliers and adaptively search for the optimal feature subspace to reduce the influence of redundant features on high-dimensional data.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Automation & Control Systems
Pierre Humbert, Batiste Le Bars, Laurent Oudre, Argyris Kalogeratos, Nicolas Vayatis
Summary: This paper introduces two algorithms, IGL-3SR and FGL-3SR, for learning graph structures from multivariate signals. These algorithms show superior performance in numerical computations and scalability compared to existing methods, with lower complexity.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Automation & Control Systems
Sichao Fu, Qiong Cao, Yunwen Lei, Yujie Zhong, Yibing Zhan, Xinge You
Summary: In this article, a novel dynamic graph structure preserving (DGSP) model is proposed for few-shot learning. The model updates the graph structure by considering the data correlations from both the feature space and the label space, effectively correcting the local geometry relationships. Extensive experiments demonstrate that the proposed method outperforms existing methods in various benchmarks, backbones, and task settings, achieving state-of-the-art performance.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Shikun Mei, Wenhui Zhao, Quanxue Gao, Ming Yang, Xinbo Gao
Summary: Recently, there has been a lot of interest in developing graph-based subspace clustering methods for high-dimensional data. These methods do not require prior knowledge of the number of dimensions and subspaces. In our proposed algorithm, we introduce feature selection to process the input data, construct a dictionary to remove redundant information, and learn the optimal bipartite graph with K-connected components. The experimental results on motion segmentation and face recognition datasets demonstrate the effectiveness and stability of our algorithm.
Article
Mathematics
Juan-Luis Garcia-Zapata, Clara Gracio
Summary: Spectral techniques based on the Laplacian matrix are often used for partitioning vertices or forming clusters in graphs. The introduction of the p-Laplacian allows for consideration of vertex weights independent of edge weights. Clustering based on the importance of vertices is enabled by assigning different weights to vertices, offering more flexibility than simply considering vertex numbers.
Article
Computer Science, Artificial Intelligence
Jun Wang, Chang Tang, Xiao Zheng, Xinwang Liu, Wei Zhang, En Zhu
Summary: This paper proposes a graph regularized spatial-spectral subspace clustering method (GRSC) for hyperspectral band selection. The method preserves the spatial information of hyperspectral images through superpixel segmentation and generates discriminative latent features to represent the bands. It explores spectral correlation using a self representation subspace clustering model and regularization, and learns a similarity graph between region-aware latent features to preserve the spatial structure of the images.
Article
Computer Science, Information Systems
Baojun Zhao, Wei Tang, Yu Pan, Yuqi Han, Wenzheng Wang
Summary: In the field of remote sensing, small inter-class and massive intra-class changes present challenges in aircraft model recognition. To address this, a novel aircraft type recognition method BD-ELMNet is proposed, integrating advantages of CNN, AE, and ELM. The method outperforms existing methods in experiments.
Article
Computer Science, Artificial Intelligence
Shifei Ding, Benyu Wu, Xiao Xu, Lili Guo, Ling Ding
Summary: This paper proposes a Graph Clustering Network with Structure Embedding Enhanced (GC-SEE) which extracts structure information based on node importance and attribute importance. It captures different orders-based structure information through multi-scale feature fusion and integrates different types of structure information using a self-supervised learning module. Experimental results on benchmark datasets demonstrate the superiority and effectiveness of GC-SEE in deep clustering.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Information Systems
Xiaoshuang Sang, Jianfeng Lu, Hong Lu
Summary: A novel consensus graph-based auto-weighted multi-view projection clustering approach is proposed in this paper, which addresses the noise and redundancy issues in multi-view data clustering through consensus structured graph learning, automatic weighting, and manifold structure preservation strategies, showing promising results.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Xue Wang, Liming Yang
Summary: Semi-supervised learning technique is proposed with the introduction of Laplacian regularization term, aiming to build two nonparallel hyperplanes in ELM feature space. Two new semi-supervised ELM classification algorithms are introduced: Laplacian Generalized Eigenvalue ELM (LapGELM) and Laplacian Standard Eigenvalue ELM (LapSELM). Experiment results demonstrate the feasibility, simplicity, rapidity, and good generalization performance of the proposed algorithms compared to traditional methods.
NEURAL PROCESSING LETTERS
(2022)
Article
Automation & Control Systems
A. Al-Ghanimi, J. Zheng, Z. Man
INTERNATIONAL JOURNAL OF CONTROL
(2017)
Article
Automation & Control Systems
Hai Wang, Liheng Shi, Zhihong Man, Jinchuan Zheng, Shihua Li, Ming Yu, Canghua Jiang, Huifang Kong, Zhenwei Cao
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2018)
Article
Automation & Control Systems
Maria Mitrevska, Zhenwei Cao, Jinchuan Zheng, Edi Kurniawan, Zhihong Man
INTERNATIONAL JOURNAL OF CONTROL AUTOMATION AND SYSTEMS
(2018)
Article
Engineering, Mechanical
Zhe Sun, Jinchuan Zheng, Zhihong Man, Minyue Fu, Renquan Lu
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2019)
Article
Automation & Control Systems
M. Mitrevska, Z. Cao, J. Zheng, E. Kurniawan, Z. Man
INTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROL
(2019)
Article
Computer Science, Artificial Intelligence
Zhenyi Shen, Zhihong Man, Zhenwei Cao, Jinchuan Zheng
NEURAL COMPUTING & APPLICATIONS
(2020)
Article
Automation & Control Systems
Jasim Khawwaf, Jinchuan Zheng, Rifai Chai, Renquan Lu, Zhihong Man
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2019)
Article
Computer Science, Artificial Intelligence
Zhe Sun, Jinchuan Zheng, Zhihong Man, Hai Wang, Ke Shao, Defeng He
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS
(2019)
Article
Automation & Control Systems
Ke Shao, Jinchuan Zheng, Rongchuan Tang, Xiu Li, Zhihong Man, Bin Liang
Summary: This article presents a barrier function based adaptive sliding mode control scheme for uncertain nonlinear systems with actuator saturation. The proposed method can adapt to time-varying disturbances and does not require the upper bound information of disturbance.
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2022)
Article
Engineering, Electrical & Electronic
Muhammad Ali Hassan, Zhenwei Cao, Zhihong Man
Summary: This research paper proposes a novel hyperbolic secant function-based sliding mode controller that achieves fast convergence, reduces chattering, and improves tracking performance. Stability analysis based on Lyapunov stability criteria is conducted. Comparative simulations and experiments on a pantograph robot illustrate the effectiveness of the proposed algorithm.
Proceedings Paper
Engineering, Electrical & Electronic
Zhenyi Shen, Zhihong Man, Zhenwei Cao, Jinchuan Zhcng
2019 INTERNATIONAL CONFERENCE ON ADVANCED MECHATRONIC SYSTEMS (ICAMECHS)
(2019)
Proceedings Paper
Engineering, Electrical & Electronic
Zhangwei Xu, Zhihong Man, Qing-Long Han, Zhenwei Cao, Jinchuan Zheng, Yew Wee Wong, Ming Huang
2019 INTERNATIONAL CONFERENCE ON ADVANCED MECHATRONIC SYSTEMS (ICAMECHS)
(2019)
Proceedings Paper
Engineering, Industrial
Linxian Zhi, Zhihong Man, Zhenwei Cao, Jinchuan Zheng
PROCEEDINGS OF THE 2019 14TH IEEE CONFERENCE ON INDUSTRIAL ELECTRONICS AND APPLICATIONS (ICIEA 2019)
(2019)
Proceedings Paper
Automation & Control Systems
Liu Linfeng, Wang Hai, He Ping, Kong Huifang, Yu Ming, Jiang Canghua, Man Zhihong
PROCEEDINGS OF THE 36TH CHINESE CONTROL CONFERENCE (CCC 2017)
(2017)
Proceedings Paper
Engineering, Industrial
Kamal Rsetam, Zhenwei Cao, Zhihong Man, Maria Mitrevska
IECON 2017 - 43RD ANNUAL CONFERENCE OF THE IEEE INDUSTRIAL ELECTRONICS SOCIETY
(2017)