Article
Computer Science, Artificial Intelligence
Zheng Wang, Haojie Hu, Rong Wang, Qianrong Zhang, Feiping Nie, Xuelong Li
Summary: A novel robust trace ratio objective is proposed in this paper to reduce the negative effects of outliers on the objective function by converting the mean calculation method, showing superior performance on several benchmark datasets.
Article
Operations Research & Management Science
Chun-Na Li, Pei-Wei Ren, Yan-Ru Guo, Ya-Fen Ye, Yuan-Hai Shao
Summary: This paper proposes a regularized linear discriminant analysis method called GCLDA, which uses a generalized capped norm to measure distances and includes a regularization term to improve adaptability and avoid singularity.
ANNALS OF OPERATIONS RESEARCH
(2022)
Article
Computer Science, Artificial Intelligence
Jiakou Liu, Xiong Xiong, Peiwei Ren, Chun-Na Li, Yuan-Hai Shao
Summary: Classical linear discriminant analysis (LDA) is sensitive to outliers and noise due to its reliance on squared Frobenious norm. To address this issue, a novel method called capped l(2,1)-norm linear discriminant analysis (CLDA) is proposed in this paper, which employs non-squared l(2)-norm and capped operation to improve robustness. Experimental results on various datasets demonstrate the effectiveness of CLDA in removing outliers and suppressing noise.
APPLIED INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Chun-Na Li, Yi-Fan Qi, Da Zhao, Tingting Guo, Lan Bai
Summary: 2DLDA is an extension of LDA that can handle matrix input samples directly. However, it is sensitive to noise and outliers. In this paper, a square-free F-norm 2DLDA is proposed to improve its robustness. By eliminating the squared operation, the proposed method weakens the influence of outliers and noise while preserving the geometric structure of data.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2022)
Article
Automation & Control Systems
Chun-Na Li, Yi-Fan Qi, Yuan-Hai Shao, Yan-Ru Guo, Ya-Fen Ye
Summary: 2DCLDA is a method to address the sensitivity of outliers and noise in 2DLDA, utilizing capped l(2,1)-norm and a regularization term to effectively identify outliers and suppress the effect of noise. It is solved through a series of generalized eigenvalue problems, with the proposed iteration algorithm monotonously decreasing the objective of 2DCLDA. Experimental results demonstrate the superiority of 2DCLDA, especially for noise data, when compared with related approaches on several face image databases.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2021)
Article
Engineering, Electrical & Electronic
Zhongliang Lyu, Hua Wei, Xiaoqing Bai, Daiyu Xie, Le Zhang, Peijie Li
Summary: This paper proposes an L-p (0 <1) quasi norm state estimator for power system static state estimation, which can effectively suppress bad data compared to existing L-1 and L-2 norm estimators. The robustness of the proposed estimator is discussed, and it is shown that its ability to suppress bad data increases with decreasing p. Moreover, an algorithm is suggested to solve the non-convex state estimation problem and prevent the solution from converging to a local optimum.
JOURNAL OF MODERN POWER SYSTEMS AND CLEAN ENERGY
(2022)
Article
Computer Science, Artificial Intelligence
Fa Zhu, Junbin Gao, Jian Yang, Ning Ye
Summary: Linear Discriminant Analysis (LDA) assumes samples from the same class are independently and identically distributed, which may lead to failure when there are multiple clusters within a class. This paper proposes a neighborhood linear discriminant analysis (nLDA) that defines scatter matrices based on a neighborhood of reverse nearest neighbors, eliminating the need for the i.i.d. assumption. Experimental results show that nLDA outperforms previous discriminators in terms of performance.
PATTERN RECOGNITION
(2022)
Article
Computer Science, Artificial Intelligence
Abhik Ghosh, Rita SahaRay, Sayan Chakrabarty, Sayan Bhadra
Summary: Quadratic discriminant analysis (QDA) and generalized quadratic discriminant analysis (GQDA) are effective statistical tools for classifying observations from different multivariate Normal populations, especially when dealing with populations with underlying elliptically symmetric distributions. However, they show significantly reduced efficiency and increased misclassification errors when handling data that are highly vulnerable to outliers.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Information Systems
Chun-Na Li, Yi Li, Yan-Hui Meng, Pei-Wei Ren, Yuan-Hai Shao
Summary: Recently, a study focused on an absolute value inequalities discriminant analysis criterion with robustness and sparseness for supervised dimensionality reduction. However, its method of obtaining discriminant directions one by one through greedy search fails to explain the sparseness of multiple discriminant directions, and it also relies on solving a series of linear programming problems, which is time-consuming. In this paper, a novel linear discriminant analysis approach is proposed, integrating robustness and sparseness using the L-1-norm and L-2, L-1-norm. The proposed method obtains all the discriminant directions simultaneously and is solved using the more efficient alternating direction method of multipliers. Experimental results on various datasets validate the effectiveness of the proposed method.
Article
Computer Science, Artificial Intelligence
Chun-Na Li, Yuan-Hai Shao, Wei-Jie Chen, Zhen Wang, Nai-Yang Deng
Summary: This paper proposes a generalized Lp-norm 2DLDA framework named G2DLDA, which can achieve robustness by selecting proper p value and improve generalization performance and avoid singularity by introducing a regularization term.
Article
Operations Research & Management Science
Chun-Na Li, Jiakou Liu, Yanhui Meng, Yuan-Hai Shao
Summary: This paper proposes a Universum linear discriminant analysis method to improve linear discriminant analysis. Compared to existing methods, this method fully utilizes Universum information to obtain discriminant directions and can obtain any number of discriminant directions.
OPTIMIZATION LETTERS
(2023)
Article
Mathematics, Interdisciplinary Applications
Jian Wang, Wenjing Jiang, Wei Shao
Summary: In this paper, a multi-fractal detrending fluctuation analysis method based on L-p-norm constraint (MF-DFN) is proposed. The performance of the algorithm is evaluated using a p-model-based multiplicative cascades time series. The results show that the appropriate norm constraints can accurately describe the multifractal characteristics of the time series, and the proposed MF-DFN method outperforms MF-DFA in analyzing the multifractal characteristics and plays a significant role in ECG classification.
FRACTALS-COMPLEX GEOMETRY PATTERNS AND SCALING IN NATURE AND SOCIETY
(2022)
Article
Computer Science, Artificial Intelligence
Rongxiu Lu, Yingjie Cai, Jianyong Zhu, Feiping Nie, Hui Yang
Summary: In this study, a new dimensionality reduction method ALDA is proposed, which effectively handles multimodal data and improves dimensionality reduction performance through automatic weighting and local structure capturing methods. Extensive experiments showed the effectiveness of ALDA compared to state-of-the-art methods.
Article
Computer Science, Information Systems
Hao Zheng, Liyong Fu, Qiaolin Ye
Summary: Robust principal component analysis (PCA) has been proven effective in data reconstruction and recognition tasks. However, existing methods often suffer from performance and robustness issues. To address this, we propose a new method called flexible capped PCA (FCPCA) that uses capped L2,p-norm distance metric to minimize reconstruction errors. Experimental results demonstrate that FCPCA outperforms existing methods in terms of power and flexibility.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Guoquan Li, Linxi Yang, Kequan Zhao
Summary: This paper proposes a unified model based on the generalized l(q)-norm to address the challenge of optimal scoring on small sample size datasets, and develops an efficient alternative direction method of multipliers to handle the difficulties in dealing with the generalized norm. Numerical experiments demonstrate the effectiveness and feasibility of the proposed method.
PATTERN RECOGNITION
(2023)
Article
Automation & Control Systems
Jie Wen, Zheng Zhang, Zhao Zhang, Lunke Fei, Meng Wang
Summary: This article introduces a simple and effective incomplete multiview clustering framework that considers incomplete multiview data, taking into account both local geometric information and unbalanced discriminating powers. By developing a novel graph-regularized matrix factorization model and introducing a semantic consistency constraint, the common representations learned from different views can maintain local geometric similarities and achieve a unified discriminative representation. Furthermore, the importance of different views is adaptively determined to reduce the negative influence of unbalanced incomplete views, leading to superior clustering performance compared to state-of-the-art multiview learning methods according to extensive experimental results on several incomplete multiview datasets.
IEEE TRANSACTIONS ON CYBERNETICS
(2021)
Review
Computer Science, Information Systems
Yuzhu Ji, Haijun Zhang, Zhao Zhang, Ming Liu
Summary: This paper investigates the application of convolutional neural network-based encoder-decoder models in the field of salient object detection. Through extensive experimental research on encoder-decoder models with different parameters, new baseline models that can outperform state-of-the-art performance were discovered.
INFORMATION SCIENCES
(2021)
Editorial Material
Computer Science, Artificial Intelligence
Zhao Zhang, Meng Wang, Sheng Li, Zheng Zhang
Article
Computer Science, Artificial Intelligence
Zhao Zhang, Yulin Sun, Yang Wang, Zheng Zhang, Haijun Zhang, Guangcan Liu, Meng Wang
Summary: The proposed SLatDPL model integrates coefficient learning and salient feature extraction, enabling simultaneous discovery of underlying subspaces and salient features. Through twin-incoherence constraint and adaptive weighting strategy, SLatDPL ensures the block-diagonality of encoding coefficients and discriminative salient features. Extensive simulations on multiple public databases demonstrate the satisfactory performance of SLatDPL compared to related methods.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Computer Science, Information Systems
Bingzhi Chen, Yishu Liu, Zheng Zhang, Yingjian Li, Zhao Zhang, Guangming Lu, Hongbing Yu
Summary: This study introduces a novel CNN framework DACE for COVID-19 diagnosis, which leverages unlabeled neighbors to progressively learn robust feature representations and generate a well-performed classifier. The framework combines a Long-Short Hierarchical Attention Network (LSHAN) and an efficient context estimation criterion to achieve superior performance.
ACM TRANSACTIONS ON MULTIMEDIA COMPUTING COMMUNICATIONS AND APPLICATIONS
(2021)
Article
Computer Science, Information Systems
Rui Gao, Xingsong Hou, Jie Qin, Yuming Shen, Yang Long, Li Liu, Zhao Zhang, Ling Shao
Summary: Zero-shot learning aims to recognize unknown categories that are not available during training. Generative models have shown potential to address this problem by synthesizing unseen features based on semantic embeddings. We propose a visual-semantic aligned bidirectional network with cycle consistency to bridge the gap between visual and semantic spaces and generate high-quality unseen features. Two carefully designed strategies are incorporated to improve the overall ZSL performance by enhancing intra-domain class divergence and mitigating inter-domain shift.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)
Article
Computer Science, Information Systems
Han Yan, Haijun Zhang, Linlin Liu, Dongliang Zhou, Xiaofei Xu, Zhao Zhang, Shuicheng Yan
Summary: The article introduces an AI-based framework for fashion design using generative adversarial networks to enhance designers' efficiency. The framework includes a sketch-generation module based on latent space and a rendering-generation module to learn the mapping between textures and sketches. Experimental results demonstrate the effectiveness of the proposed method in synthesizing semantic-aware textures on sketches.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)
Article
Computer Science, Artificial Intelligence
Hongwei Yin, Guixiang Wang, Wenjun Hu, Zhao Zhang
Summary: Multi-view clustering is a hot research topic that leverages complementary information from multiple views to improve clustering performance. In this paper, a novel fine-grained multi-view clustering method is proposed. It divides the sample space of each view into sub-clusters using multi-prototypes representation, enhances the robustness of representation by reducing sub-cluster overlap, and assigns contribution weights based on clustering capacity. The method integrates robust multi-prototypes representation, fine-grained multi-view fusion, and clustering process into a unified framework, and achieves better clustering accuracy compared to traditional methods.
APPLIED INTELLIGENCE
(2023)
Article
Automation & Control Systems
Qiaolin Ye, Peng Huang, Zhao Zhang, Yuhui Zheng, Liyong Fu, Wankou Yang
Summary: This article presents a new multiview learning approach, MvRDTSVM, to improve classification performance and robustness by introducing double-sided constraints and using L1-norm as the distance metric. Experimental results confirm the effectiveness of the proposed methods.
IEEE TRANSACTIONS ON CYBERNETICS
(2022)
Article
Automation & Control Systems
Jie Wen, Zheng Zhang, Lunke Fei, Bob Zhang, Yong Xu, Zhao Zhang, Jinxing Li
Summary: Conventional multiview clustering methods fail when not all views of samples are available in practical applications. Incomplete multiview clustering (IMC) is developed to address this issue. Recent years have seen significant advances in IMC research, but there are still open problems to be solved.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Lei Ma, Yuhui Zheng, Zhao Zhang, Yazhou Yao, Xijian Fan, Qiaolin Ye
Summary: Conventional action recognition algorithms face great challenges in recognizing unseen combinations of action and different objects, which is known as (zero-shot) compositional action recognition in real-world applications. Previous methods rely heavily on manual annotation or the quality of detectors to enhance the dynamic clues of objects in the scene. In this work, we propose a novel Motion Stimulation (MS) block to autonomously mine the temporal clues from moving objects or hands without explicit supervision, which can enhance the ability of compositional generalization for action recognition algorithms when integrated into existing video backbones. Experimental results on three action recognition datasets demonstrate the effectiveness and interpretability of our MS block.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Business
Yuan Sun, Zhebin Ding, Zuopeng Zhang
Summary: In recent years, China has made great progress in the adoption and use of emerging information technologies. The popularity of enterprise social media (ESM) in Chinese enterprises and the resulting innovation practices provide an opportunity to explore relevant theories and assumptions. This article applies the organizational information processing theory (OIPT) to study the innovative use cases of ESM in China and identify contributing factors through a multicase-analysis approach. The research contributes to the literature by studying the innovative use cases in the ESM context and provides valuable insights for practitioners in designing and implementing ESM effectively.
IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT
(2023)
Article
Computer Science, Artificial Intelligence
Dongliang Zhou, Haijun Zhang, Kai Yang, Linlin Liu, Han Yan, Xiaofei Xu, Zhao Zhang, Shuicheng Yan
Summary: The article introduces a novel outfit generation framework OutfitGAN, aiming to synthesize a set of complementary items to compose an entire outfit. Through extensive experiments on a large-scale dataset, OutfitGAN demonstrates superior performance in synthesizing photo-realistic outfits and improving compatibility.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Jie Wen, Shijie Deng, Lunke Fei, Zheng Zhang, Bob Zhang, Zhao Zhang, Yong Xu
Summary: In this article, a new linear regression-based multiclass classification method, called DRAGD, is proposed. It explores the high-order structure information and provides a new way to capture the structure of data, resulting in a more discriminative transformation matrix. Experimental results show that DRAGD outperforms existing LR methods.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Yanyan Wei, Zhao Zhang, Yang Wang, Mingliang Xu, Yi Yang, Shuicheng Yan, Meng Wang
Summary: The paper introduces a new unsupervised framework, DerainCycleGAN, for single image rain removal and generation, which utilizes transfer learning and cyclic structures effectively. It also presents an unsupervised rain attentive detector (UARD) to enhance rain information detection, and a new synthetic method for generating rain streaks, improving the processing of real rainy images.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2021)
Article
Computer Science, Artificial Intelligence
Hamdan Abdellatef, Lina J. Karam
Summary: This paper proposes performing the learning and inference processes in the compressed domain to reduce computational complexity and improve speed of neural networks. Experimental results show that modified ResNet-50 in the compressed domain is 70% faster than traditional spatial-based ResNet-50 while maintaining similar accuracy. Additionally, a preprocessing step with partial encoding is suggested to improve resilience to distortions caused by low-quality encoded images. Training a network with highly compressed data can achieve good classification accuracy with significantly reduced storage requirements.
Article
Computer Science, Artificial Intelligence
Victor R. Barradas, Yasuharu Koike, Nicolas Schweighofer
Summary: Inverse models are essential for human motor learning as they map desired actions to motor commands. The shape of the error surface and the distribution of targets in a task play a crucial role in determining the speed of learning.
Article
Computer Science, Artificial Intelligence
Ting Zhou, Hanshu Yan, Jingfeng Zhang, Lei Liu, Bo Han
Summary: We propose a defense strategy that reduces the success rate of data poisoning attacks in downstream tasks by pre-training a robust foundation model.
Article
Computer Science, Artificial Intelligence
Hao Sun, Li Shen, Qihuang Zhong, Liang Ding, Shixiang Chen, Jingwei Sun, Jing Li, Guangzhong Sun, Dacheng Tao
Summary: In this paper, the convergence rate of AdaSAM in the stochastic non-convex setting is analyzed. Theoretical proof shows that AdaSAM has a linear speedup property and decouples the stochastic gradient steps with the adaptive learning rate and perturbed gradient. Experimental results demonstrate that AdaSAM outperforms other optimizers in terms of performance.
Article
Computer Science, Artificial Intelligence
Juntong Yun, Du Jiang, Li Huang, Bo Tao, Shangchun Liao, Ying Liu, Xin Liu, Gongfa Li, Disi Chen, Baojia Chen
Summary: In this study, a dual manipulator grasping detection model based on the Markov decision process is proposed. By parameterizing the grasping detection model of dual manipulators using a cross entropy convolutional neural network and a full convolutional neural network, stable grasping of complex multiple objects is achieved. Robot grasping experiments were conducted to verify the feasibility and superiority of this method.
Article
Computer Science, Artificial Intelligence
Miaohui Zhang, Kaifang Li, Jianxin Ma, Xile Wang
Summary: This paper proposes an unsupervised person re-identification (Re-ID) method that uses two asymmetric networks to generate pseudo-labels for each other by clustering and updates and optimizes the pseudo-labels through alternate training. It also designs similarity compensation and similarity suppression based on the camera ID of pedestrian images to optimize the similarity measure. Extensive experiments show that the proposed method achieves superior performance compared to state-of-the-art unsupervised person re-identification methods.
Article
Computer Science, Artificial Intelligence
Florian Bacho, Dominique Chu
Summary: This paper proposes a new approach called the Forward Direct Feedback Alignment algorithm for supervised learning in deep neural networks. By combining activity-perturbed forward gradients, direct feedback alignment, and momentum, this method achieves better performance and convergence speed compared to other local alternatives to backpropagation.
Article
Computer Science, Artificial Intelligence
Xiaojian Ding, Yi Li, Shilin Chen
Summary: This research paper addresses the limitations of recursive feature elimination (RFE) and its variants in high-dimensional feature selection tasks. The proposed algorithms, which introduce a novel feature ranking criterion and an optimal feature subset evaluation algorithm, outperform current state-of-the-art methods.
Article
Computer Science, Artificial Intelligence
Naoko Koide-Majima, Shinji Nishimoto, Kei Majima
Summary: Visual images observed by humans can be reconstructed from brain activity, and the visualization of arbitrary natural images from mental imagery has been achieved through an improved method. This study provides a unique tool for directly investigating the subjective contents of the brain.
Article
Computer Science, Artificial Intelligence
Huanjie Tao, Qianyue Duan
Summary: In this paper, a hierarchical attention network with progressive feature fusion is proposed for facial expression recognition (FER), addressing the challenges posed by pose variation, occlusions, and illumination variation. The model achieves enhanced performance by aggregating diverse features and progressively enhancing discriminative features.
Article
Computer Science, Artificial Intelligence
Zhenyi Wang, Pengfei Yang, Linwei Hu, Bowen Zhang, Chengmin Lin, Wenkai Lv, Quan Wang
Summary: In the face of the complex landscape of deep learning, we propose a novel subgraph-level performance prediction method called SLAPP, which combines graph and operator features through an innovative graph neural network called EAGAT, providing accurate performance predictions. In addition, we introduce a mixed loss design with dynamic weight adjustment to improve predictive accuracy.
Article
Computer Science, Artificial Intelligence
Yiyang Yin, Shuangling Luo, Jun Zhou, Liang Kang, Calvin Yu-Chian Chen
Summary: Medical image segmentation is crucial for modern healthcare systems, especially in reducing surgical risks and planning treatments. Transanal total mesorectal excision (TaTME) has become an important method for treating colon and rectum cancers. Real-time instance segmentation during TaTME surgeries can assist surgeons in minimizing risks. However, the dynamic variations in TaTME images pose challenges for accurate instance segmentation.
Article
Computer Science, Artificial Intelligence
Teng Cheng, Lei Sun, Junning Zhang, Jinling Wang, Zhanyang Wei
Summary: This study proposes a scheme that combines the start-stop point signal features for wideband multi-signal detection, called Fast Spectrum-Size Self-Training network (FSSNet). By utilizing start-stop points to build the signal model, this method successfully solves the difficulty of existing deep learning methods in detecting discontinuous signals and achieves satisfactory detection speed.
Article
Computer Science, Artificial Intelligence
Wenming Wu, Xiaoke Ma, Quan Wang, Maoguo Gong, Quanxue Gao
Summary: The layer-specific modules in multi-layer networks are critical for understanding the structure and function of the system. However, existing methods fail to accurately characterize and balance the connectivity and specificity of these modules. To address this issue, a joint learning graph clustering algorithm (DRDF) is proposed, which learns the deep representation and discriminative features of the multi-layer network, and balances the connectivity and specificity of the layer-specific modules through joint learning.
Article
Computer Science, Artificial Intelligence
Guanghui Yue, Guibin Zhuo, Weiqing Yan, Tianwei Zhou, Chang Tang, Peng Yang, Tianfu Wang
Summary: This paper proposes a novel boundary uncertainty aware network (BUNet) for precise and robust colorectal polyp segmentation. BUNet utilizes a pyramid vision transformer encoder to learn multi-scale features and incorporates a boundary exploration module (BEM) and a boundary uncertainty aware module (BUM) to handle boundary areas. Experimental results demonstrate that BUNet outperforms other methods in terms of performance and generalization ability.