Article
Computer Science, Hardware & Architecture
Hao Liang, Li Kang, Jianjun Huang
Summary: The low-rank matrix completion problem has attracted attention in engineering and applied sciences. Existing methods have limitations, and this paper proposes a robust matrix completion model that overcomes these limitations. Numerical simulations and experiments demonstrate the effectiveness and advantages of the proposed method.
JOURNAL OF SUPERCOMPUTING
(2022)
Article
Computer Science, Artificial Intelligence
Linxi Yang, Guoquan Li, Zhiyou Wu, Changzhi Wu
Summary: This paper introduces a new robust truncated L-2-norm twin support vector machine, which uses truncated L-2-norm to measure empirical risk and employs chance constraints to specify error rates. Experimental results demonstrate the significant advantages of this method in robustness and generalization performance.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2021)
Article
Computer Science, Artificial Intelligence
Wei Jiang, Jun Zhang, Changsheng Zhang, Lijun Wang, Heng Qi
Summary: Recent research has shown that low tubal rank recovery based on tensor has attracted attention. This study proposes two novel tractable tensor completion models and defines tensor double nuclear norm, tensor Frobenius/nuclear hybrid norm as proxies for tensor tubal rank. Experimental results demonstrate better performance compared to some state-of-the-art methods.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Artificial Intelligence
Chenglin Wen, Wenchao Qian, Qinghua Zhang, Feilong Cao
Summary: The paper introduces matrix completion and low-rank sparse decomposition models based on truncated Schatten p-norm, which are more flexible than nuclear norm. By transforming non-convex optimization models into convex ones using function expansion method, and solving them using the ADMM-based two-step iterative algorithm, the convergence of the proposed algorithm is mathematically proven. The superiority of the proposed method is further confirmed by comparing it with existing methods on synthetic data and actual images.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2021)
Article
Engineering, Multidisciplinary
Yuki Kondo, Munetoshi Numada, Ichiro Yoshida, Yudai Yamaguchi, Hirokazu Machida, Hiroyasu Koshimizu
Summary: In this paper, a novel robust filter based on L2-norm, L2GF, is proposed, which exhibits higher robustness and processing speed.
PRECISION ENGINEERING-JOURNAL OF THE INTERNATIONAL SOCIETIES FOR PRECISION ENGINEERING AND NANOTECHNOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Jiaqi Bao, Mineichi Kudo, Keigo Kimura, Lu Sun
Summary: A novel approach called robust embedding regression (RER) for semi-supervised learning is proposed in this paper, which constructs a more robust and accurate graph by inheriting the advantages of existing techniques and introduces low-rank representation to reduce the negative influence of redundant features and noises. Experimental results demonstrate that RER outperforms existing methods in classification and clustering performance.
PATTERN RECOGNITION
(2024)
Article
Engineering, Multidisciplinary
Yuki Kondo, Munetoshi Numada, Ichiro Yoshida, Yudai Yamaguchi, Hirokazu Machida, Hiroyasu Koshimizu
Summary: A new robust filter called L1-norm Gaussian filter (L1GF) is proposed in this research, which behaves robustly to spikes, steps, and slopes as examples given in ISO16610-30. This solves the inconsistency problem between ISO standards.
Article
Computer Science, Information Systems
Yanmeng Li, Huaijiang Sun, Wenzhu Yan, Qiongjie Cui
Summary: In this study, a novel Robust Capped L-1-norm Twin Support Vector Machine with Privileged Information (R-CTSVM+) is proposed to achieve better performance in the presence of noise and outliers in the data. The pair of regularization functions designed in the model increases the model's tolerance to disturbances, while the capped L-1 regularized distance ensures the robustness of the model. Experimental results demonstrate the superiority of the proposed model in handling data with a significant amount of noise and outliers.
INFORMATION SCIENCES
(2021)
Article
Chemistry, Analytical
Hong Xia, Qingyi Dong, Jiahao Zheng, Yanping Chen, Cong Gao, Zhongmin Wang
Summary: This paper proposes a truncated nuclear norm low-rank tensor completion method for QoS data prediction, which improves prediction accuracy by constructing tensors and introducing truncated nuclear norm.
Article
Computer Science, Artificial Intelligence
Ting Shi, Sugen Chen
Summary: In this paper, a novel robust twin support vector regression algorithm TH epsilon-TSVR is proposed to handle noise and outliers in regression problems. A smooth truncated H epsilon loss function is constructed by combining epsilon-insensitive loss and Huber loss, and a concave-convex programming method is used to solve the nonconvex optimization problem in the primal space. The experimental results verify the effectiveness and robustness of TH epsilon-TSVR.
NEURAL PROCESSING LETTERS
(2023)
Article
Engineering, Electrical & Electronic
Zhi-Yong Wang, Xiao Peng Li, Hing Cheung So
Summary: Robust matrix completion aims to recover a low-rank matrix from a subset of entries corrupted by gross errors. Most existing methods perform well for noise-free data or data with zero-mean white Gaussian noise, but their performance degrades in the presence of outliers. In this paper, we propose a new robust matrix completion scheme based on the factorization framework, using the truncated-quadratic loss function and half-quadratic optimization. Our algorithms outperform state-of-the-art methods in terms of restoration accuracy and runtime, as demonstrated by numerical simulations and experimental results on image inpainting and hyperspectral image recovery. MATLAB code is available at https://github.com/bestzywang.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Automation & Control Systems
Junhong Zhang, Zhihui Lai, Heng Kong, Linlin Shen
Summary: In this paper, a new robust manifold twin bounded SVM (RMTBSVM) algorithm is proposed, which considers both robustness and discriminability. By using the capped L-1-norm as the distance metric and adding robust manifold regularization, the robustness and classification performance are improved. The algorithm is extended for nonlinear classification using the kernel method.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Computer Science, Information Systems
Liqiao Yang, Kit Ian Kou, Jifei Miao
Summary: In recent years, quaternion matrix completion based on low-rank regularization has been increasingly utilized in image processing, providing a more comprehensive approach for handling RGB images compared to traditional low-rank matrix completion. The introduction of quaternion truncated nuclear norm allows for a more accurate approximation of low-rank attributes, while the weighted method and concise gradient descent strategy accelerates the convergence of the QTNN method with theoretical guarantees in optimization.
JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION
(2021)
Article
Computer Science, Artificial Intelligence
Yanmeng Li, Huaijiang Sun
Summary: In this paper, a new method named Safe Sample Screening for robust TSVM (SSS-RTSVM) is proposed. SSS-RTSVM clips the hinge loss in the traditional soft margin twin support vector machine to the ramp loss, and provides a pair of nonparallel proximal hyperplanes to achieve good anti-noise ability. Additionally, safe sample screening rules based on CCCP are integrated to reduce the computational cost without sacrificing the optimal accuracy.
APPLIED INTELLIGENCE
(2023)
Article
Computer Science, Information Systems
Qiang Li, Ziqi Xie, Lihong Wang
Summary: The subspace clustering algorithm by block diagonal representation (BDR) is effective for subspace clustering on noise-free datasets drawn from independent linear subspaces. However, real-world data are usually corrupted by noises and have overlapping subspaces, causing degraded performance of linear subspace clustering algorithms, including BDR. To address this issue, we propose a new objective function based on BDR, which incorporates the l(2,1) norm of the reconstruction error to model noises and enhance algorithm robustness. We present a corresponding subspace clustering algorithm to handle noisy datasets, where a coefficient matrix with a block diagonal structure is pursued and used to construct an affinity matrix for spectral clustering. Experimental results on artificial noisy image datasets demonstrate the robustness and improved clustering performance of the proposed algorithm compared to other methods.
Article
Engineering, Mechanical
Jinde Zheng, Haiyang Pan
NONLINEAR DYNAMICS
(2020)
Article
Automation & Control Systems
Jinde Zheng, Haiyang Pan, Jinyu Tong, Qingyun Liu
Summary: Extracting failure-related information from vibration signals is crucial for vibration-based fault detection in rolling bearings. This article proposes a new nonlinear dynamic parameter to enhance the measurement of data complexity and compares it with existing algorithms. Furthermore, a novel fault diagnosis approach is introduced, which achieves the highest identifying rate and the best performance among the comparative approaches.
Article
Computer Science, Artificial Intelligence
Mingen Gu, Jinde Zheng, Haiyang Pan, Jinyu Tong
Summary: The paper introduces a new matrix classification method RSSMM, which solves the problems associated with traditional SMM by limiting the loss threshold, introducing the generalized forward-backward algorithm, and designing a generalized smooth Ramp loss function, achieving superior results in the classification of roller bearing fault signals.
APPLIED SOFT COMPUTING
(2021)
Article
Engineering, Electrical & Electronic
Wanming Ying, Jinde Zheng, Haiyang Pan, Qingyun Liu
Summary: PEUPEMD method improves the decomposition effect of UPEMD by adding uniform phase sinusoidal signals as masking signals, addressing issues such as noise residue and incomplete decomposition. Experimental results demonstrate that PEUPEMD outperforms other comparative methods in terms of decomposition accuracy and mode mixing suppression.
DIGITAL SIGNAL PROCESSING
(2021)
Article
Automation & Control Systems
Jinde Zheng, Shijun Cao, Haiyang Pan, Qing Ni
Summary: This paper proposes a novel spectral envelope-based adaptive empirical Fourier decomposition (SEAEFD) method to improve the performance of AEFD in rolling bearing vibration signal analysis. SEAEFD optimizes the spectrum segmentation boundary to achieve adaptive segmentation and minimize noise components, allowing nonstationary signals to be decomposed into single-component signals with physical significance.
Article
Computer Science, Artificial Intelligence
Haiyang Pan, Li Sheng, Haifeng Xu, Jinyu Tong, Jinde Zheng, Qingyun Liu
Summary: This paper introduces support matrix machine (SMM) as a classical matrix classification technology, and proposes a new method called pinball transfer support matrix machine (PTSMM) to solve the issue of insufficient annotation samples in practical industrial practice. The experimental results show that PTSMM effectively utilizes samples from source and target domains for modeling, and achieves higher diagnostic accuracy compared to SMM and its improved algorithms.
APPLIED SOFT COMPUTING
(2022)
Article
Materials Science, Multidisciplinary
Jinyu Tong, Cang Liu, Haiyang Pan, Jinde Zheng
Summary: This paper proposes a multisensor feature fusion method to improve the diagnosis accuracy of rolling bearings. By utilizing deep autoencoder network and variational mode decomposition, the features from multiple sensors are fused, resulting in enhanced classification accuracy for bearing fault diagnosis.
Editorial Material
Materials Science, Multidisciplinary
Ke Feng, Qing Ni, Jinde Zheng
Article
Engineering, Electrical & Electronic
Jinyu Tong, Cang Liu, Jiahan Bao, Haiyang Pan, Jinde Zheng
Summary: This article proposes a novel ensemble learning-based multisensor information fusion method to improve the effect of multisensor information fusion, reduce the discrepancy between the real value and predicted value, and enhance the accuracy of rolling bearing fault diagnosis. A multiscale convolutional neural network (MSCNN) is constructed as a base learning model to learn the multiscale features of raw vibration signals. Based on the ensemble learning framework, a multibranch MSCNN is built to extract multiple sensor signal features simultaneously and output the decision score of each sensor. The proposed method is validated on two different types of rolling bearing datasets.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Engineering, Mechanical
Jiaqi Li, Jinde Zheng, Haiyang Pan, Jinyu Tong, Ke Feng, Qing Ni
Summary: This paper proposes a new two-dimensional multi-scale time-frequency reverse dispersion entropy (CMTFRDE2D) algorithm to analyze the complexity characteristics of vibration signals, improving the instability of MDE1D. By introducing a composite coarse-grained process, the CMTFRDE2D algorithm can preserve more useful information. Experimental results show that this method can successfully extract fault information from rolling bearing vibration signals in the time-frequency domain and accurately identify different fault locations and severities of rolling bearings.
NONLINEAR DYNAMICS
(2023)
Article
Engineering, Electrical & Electronic
Shijun Cao, Jinde Zheng, Guoliang Peng, Haiyang Pan, Ke Feng, Qing Ni
Summary: The multivariate EAEFD (MEAEFD) approach is proposed in this article to deal with the mode separation problem of multichannel signals for rolling bearings and realize the self-adaptive synchronous analysis of multivariate signals. The MEAEFD-based mechanical fault diagnosis method is further proposed by fusing the multichannel feature information. The experimental results show that the MEAEFD method has advantages in decomposition accuracy and robustness, and the proposed approach has better diagnostic accuracy compared to the other methods.
IEEE SENSORS JOURNAL
(2023)
Article
Computer Science, Information Systems
Haiyang Pan, Wanwan Jiang, Qingyun Liu, Jinde Zheng
Summary: This paper proposes a multivariate nonlinear sparse mode decomposition method based on nonlinear sparse mode decomposition, which constrains singular local linear operators to separate the natural oscillation modes in multi-channel signals, showing advantages in suppressing mode aliasing and robustness.
Article
Engineering, Electrical & Electronic
Jinde Zheng, Xinglong Wang, Haiyang Pan, Jinyu Tong, Jun Zhang, Qingyun Liu
Summary: Resonance demodulation is a commonly used method for obtaining fault information, with the main challenge being to find a suitable frequency band for demodulation. Autogram is a method recently proposed for optimal demodulation frequency band selection utilizing a binary tree structure. However, the frequency band information obtained by Autogram can be easily missed. To address this issue, a 1/3 binary tree structure is used to segment the frequency domain and improve Autogram's segmentation accuracy. Despite these improvements, the method proposed in this article, TSCgram, overcomes the shortcomings of Autogram and offers better fault diagnostic results for rolling bearing analysis.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Article
Computer Science, Information Systems
Jinde Zheng, Mingen Gu, Haiyang Pan, Jinyu Tong
Article
Computer Science, Artificial Intelligence
Hao Yang, Min Wang, Zhengfei Yu, Hang Zhang, Jinshen Jiang, Yun Zhou
Summary: In this paper, a novel method called CSTTA is proposed for test time adaptation (TTA), which utilizes confidence-based optimization and sample reweighting to better utilize sample information. Extensive experiments demonstrate the effectiveness of the proposed method.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Jin Liu, Ju-Sheng Mi, Dong-Yun Niu
Summary: This article focuses on a novel method for generating a canonical basis for decision implications based on object-induced operators (OE operators). The logic of decision implication based on OE operators is described, and a method for obtaining the canonical basis for decision implications is given. The completeness, nonredundancy, and optimality of the canonical basis are proven. Additionally, a method for generating true premises based on OE operators is proposed.
KNOWLEDGE-BASED SYSTEMS
(2024)
Review
Computer Science, Artificial Intelligence
Kun Bu, Yuanchao Liu, Xiaolong Ju
Summary: This paper discusses the importance of sentiment analysis and pre-trained models in natural language processing, and explores the application of prompt learning. The research shows that prompt learning is more suitable for sentiment analysis tasks and can achieve good performance.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Xiangjun Cai, Dagang Li
Summary: This paper presents a new decomposition mechanism based on learned decomposition mapping. By using a neural network to learn the relationship between original time series and decomposed results, the repetitive computation overhead during rolling decomposition is relieved. Additionally, extended mapping and partial decomposition methods are proposed to alleviate boundary effects on prediction performance. Comparative studies demonstrate that the proposed method outperforms existing RDEMs in terms of operation speed and prediction accuracy.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Xu Wu, Yang Liu, Jie Tian, Yuanpeng Li
Summary: This paper proposes a blockchain-based privacy-preserving trust management architecture, which adopts federated learning to train task-specific trust models and utilizes differential privacy to protect device privacy. In addition, a game theory-based incentive mechanism and a parallel consensus protocol are proposed to improve the accuracy of trust computing and the efficiency of consensus.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Zaiyang Yu, Prayag Tiwari, Luyang Hou, Lusi Li, Weijun Li, Limin Jiang, Xin Ning
Summary: This study introduces a 3D view-based approach that effectively handles occlusions and leverages the geometric information of 3D objects. The proposed method achieves state-of-the-art results on occluded ReID tasks and exhibits competitive performance on holistic ReID tasks.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Yongliang Shi, Runyi Yang, Zirui Wu, Pengfei Li, Caiyun Liu, Hao Zhao, Guyue Zhou
Summary: Neural implicit representations have gained attention due to their expressive, continuous, and compact properties. However, there is still a lack of research on city-scale continual implicit dense mapping based on sparse LiDAR input. In this study, a city-scale continual neural mapping system with a panoptic representation is developed, incorporating environment-level and instance-level modeling. A tailored three-layer sampling strategy and category-specific prior are proposed to address the challenges of representing geometric information in city-scale space and achieving high fidelity mapping of instances under incomplete observation.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Ruihan Hu, Zhi-Ri Tang, Rui Yang, Zhongjie Wang
Summary: Mesh data is crucial for 3D computer vision applications worldwide, but traditional deep learning frameworks have struggled with handling meshes. This paper proposes MDSSN, a simple mesh computation framework that models triangle meshes and represents their shape using face-based and edge-based Riemannian graphs. The framework incorporates end-to-end operators inspired by traditional deep learning frameworks, and includes dedicated modules for addressing challenges in mesh classification and segmentation tasks. Experimental results demonstrate that MDSSN outperforms other state-of-the-art approaches.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Buliao Huang, Yunhui Zhu, Muhammad Usman, Huanhuan Chen
Summary: This paper proposes a novel semi-supervised conditional normalizing flow (SSCFlow) algorithm that combines unsupervised imputation and supervised classification. By estimating the conditional distribution of incomplete instances, SSCFlow facilitates imputation and classification simultaneously, addressing the issue of separated tasks ignoring data distribution and label information in traditional methods.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Deeksha Varshney, Asif Ekbal, Erik Cambria
Summary: This paper focuses on the neural-based interactive dialogue system that aims to engage and retain humans in long-lasting conversations. It proposes a new neural generative model that combines step-wise co-attention, self-attention-based transformer network, and an emotion classifier to control emotion and knowledge transfer during response generation. The results from quantitative, qualitative, and human evaluation show that the proposed models can generate natural and coherent sentences, capturing essential facts with significant improvement over emotional content.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Junchen Ye, Weimiao Li, Zhixin Zhang, Tongyu Zhu, Leilei Sun, Bowen Du
Summary: Modeling multivariate time series has long been a topic of interest for scholars in various fields. This paper introduces MvTS, an open library based on Pytorch, which provides a unified framework for implementing and evaluating these models. Extensive experiments on public datasets demonstrate the effectiveness and universality of the models reproduced by MvTS.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Reham R. Mostafa, Ahmed M. Khedr, Zaher Al Aghbari, Imad Afyouni, Ibrahim Kamel, Naveed Ahmed
Summary: Feature selection is crucial in classification procedures, but it faces challenges in high-dimensional datasets. To overcome these challenges, this study proposes an Adaptive Hybrid-Mutated Differential Evolution method that incorporates the mechanics of the Spider Wasp Optimization algorithm and the concept of Enhanced Solution Quality. Experimental results demonstrate the effectiveness of the method in terms of accuracy and convergence speed, and it outperforms contemporary cutting-edge algorithms.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Ti Xiang, Pin Lv, Liguo Sun, Yipu Yang, Jiuwu Hao
Summary: This paper introduces a Track Classification Model (TCM) based on marine radar, which can effectively recognize and classify shipping tracks. By using a feature extraction network with multi-feature fusion and a dataset production method to address missing labels, the classification accuracy is improved, resulting in successful engineering application in real scenarios.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Zhihao Zhang, Yuan Zuo, Chenghua Lin, Junjie Wu
Summary: This paper proposes a novel unsupervised context-aware quality phrase mining framework called LMPhrase, which is built upon large pre-trained language models. The framework mines quality phrases as silver labels using a parameter-free probing technique on the pre-trained language model BERT, and formalizes the phrase tagging task as a sequence generation problem by fine-tuning on the Sequence to-Sequence pre-trained language model BART. The results of extensive experiments show that LMPhrase consistently outperforms existing competitors in two different granularity phrase mining tasks.
KNOWLEDGE-BASED SYSTEMS
(2024)
Article
Computer Science, Artificial Intelligence
Kemal Buyukkaya, M. Ozan Karsavuran, Cevdet Aykanat
Summary: The study aims to investigate the hybrid parallelization of the Stochastic Gradient Descent (SGD) algorithm for solving the matrix completion problem on a high-performance computing platform. A hybrid parallel decentralized SGD framework with asynchronous inter-process communication and a novel flexible partitioning scheme is proposed to achieve scalability up to hundreds of processors. Experimental results on real-world benchmark datasets show that the proposed algorithm achieves 6x higher throughput on sparse datasets compared to the state-of-the-art, while achieving comparable throughput on relatively dense datasets.
KNOWLEDGE-BASED SYSTEMS
(2024)