Article
Mathematics, Applied
Xiaohui Chen, Yun Yang
Summary: Diffusion K-means is a clustering method suitable for data with non-linear and non-Euclidean geometric features in mixed dimensions. By proposing a polynomial-time convex relaxation algorithm via semidefinite programming, exact recovery of the SDPs for diffusion K-means is achievable under certain conditions.
APPLIED AND COMPUTATIONAL HARMONIC ANALYSIS
(2021)
Article
Computer Science, Information Systems
Haiyao Dong, Haoming Ma, Zhenguang Du, Zhicheng Zhou, Haitao Yang, Zhenyuan Wang
Summary: This study proposes a novel method called DSRS to address the limitations of current dynamic graph methods in capturing high-order global information. By using random walk techniques to mine high-order structural information, DSRS aims to balance the dynamic and high-order structures. Experimental results demonstrate significant improvements in link prediction compared to existing methods, and sensitivity testing and ablation experiments confirm the effectiveness of the proposed pre-training and parameter fine-tuning methods.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2023)
Article
Computer Science, Interdisciplinary Applications
Zahid Halim, Hussain Mahmood Sargana, Aadam, Uzma, Muhammad Waqas
Summary: This paper introduces a random walk-based method to cluster graphs, which uses information of nodes and edges to guide the random walk process and achieve clustering by finding weighted edges and neighboring nodes. Experimental results suggest better performance of this method on evaluation metrics across 18 real-world benchmark datasets.
JOURNAL OF COMPUTATIONAL SCIENCE
(2021)
Article
Mathematics
Peng Zhao, Hongjie Wu, Shudong Huang
Summary: This paper explores the implied adaptive manifold for multi-view graph clustering. By seamlessly integrating multiple adaptive graphs into a consensus graph and using a rank constraint, our model can achieve a discrete clustering result directly. In experiments, our method performs the best in terms of several evaluation metrics, demonstrating the effectiveness of the proposed approach. Meanwhile, the proposed algorithm also exhibits efficient computational performance.
Article
Computer Science, Artificial Intelligence
Chenhui Gao, Wenzhi Chen, Feiping Nie, Weizhong Yu, Feihu Yan
Summary: In this paper, we propose two algorithms, FDKM and IFDKM, for clustering high-dimensional data in a low-dimensional subspace. These algorithms have higher efficiency and lower time complexity compared to traditional methods, and their superior performance is demonstrated in multiple experiments.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Luc Giffon, Valentin Emiya, Hachem Kadri, Liva Ralaivola
Summary: K-means algorithm and Lloyd's algorithm have expanded beyond their original clustering purposes to play pivotal roles in various machine learning and data analysis techniques. QuicK-means is an efficient extension of K-means that reduces computational complexity through sparse matrix products, demonstrating benefits through experimental results.
Article
Computer Science, Artificial Intelligence
Peng Zhou, Liang Du, Xuejun Li
Summary: This paper proposes a method to improve the performance of consensus clustering by refining the base results and optimizing the ensemble process. The proposed method outperforms both single clustering and state-of-the-art consensus clustering methods.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Physics, Multidisciplinary
Daohua Yu, Xin Zhou, Yu Pan, Zhendong Niu, Huafei Sun
Summary: With the globalization of higher education, academic evaluation has gained increasing significance. This paper investigates the evaluation of academic performance using the statistical K-means (SKM) algorithm and mapping evaluation data from Euclidean space to Riemannian space, resulting in accurate clustering results. The simulation results demonstrate the advantages of the SKM algorithm.
Article
Computer Science, Artificial Intelligence
Ben Yang, Xuetao Zhang, Zhongheng Li, Feiping Nie, Fei Wang
Summary: Multi-view clustering has attracted attention for its ability to integrate information from distinct views, but improving efficiency remains a hot research topic. Anchor graph-based methods and k-means-based methods are efficient approaches but have limitations. To address these issues, we developed EMKMC, an efficient multi-view k-means clustering method that integrates anchor graphs and k-means strategies. Extensive experiments show that EMKMC significantly improves clustering efficiency without sacrificing effectiveness.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Xueying Zhu, Jie Sun, Zhenhao He, Jiantong Jiang, Zeke Wang
Summary: K-means is widely used for clustering due to its simplicity and high quality. However, it suffers from high computational complexity. To reduce costs, a mini-batch version called mbatch K-means was proposed. Although it converges faster, it introduces staleness and decreases convergence quality. In this article, a staleness-reduction version called srmbatch K-means is proposed, which achieves both low costs and high quality. Experimental results show that srmbatch can converge 40-130 times faster than mbatch, with a lower final loss.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Han Lu, Quanxue Gao, Xiangdong Zhang, Wei Xia
Summary: This study proposes a multi-view clustering framework based on K-means and graph, and develops an efficient clustering method. The method calculates the discrete cluster assignment matrix directly and considers the different contributions of views using an auto-weighted strategy. It fits both Gaussian and non-Gaussian distributed datasets.
Article
Computer Science, Artificial Intelligence
Amir Ahmad, Shehroz S. Khan
Summary: A novel algorithm initKmix is proposed for finding an initial partition for mixed datasets in k-means-based clustering algorithms. By combining clustering results from multiple runs, initKmix consistently produces accurate results and outperforms random partitioning and other initialization methods. Experimental results show that k-means clustering with initKmix performs similarly or better than state-of-the-art clustering algorithms for categorical and mixed datasets.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Jayasree Saha, Jayanta Mukherjee
Summary: The proposed algorithm in this paper can learn the number of clusters and perform clustering by comparing the distribution of a large-sized randomly sampled dataset with the original data. This scalable solution was developed based on observations regarding the similarity between sampled and original data distributions, as well as the approximation of cluster centroids. The algorithm showed significant improvements in speed and quality for predicting cluster numbers and composition in large datasets.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Artificial Intelligence
Shuqin Wang, Yongyong Chen, Shuang Yi, Guoqing Chao
Summary: In this study, a novel multi-view subspace clustering method named Frobenius norm-regularized robust graph learning (RGL) is proposed. The proposed method addresses the challenges of traditional multi-view clustering methods by exploring global information and local similarity, and removing noise and outliers using l(2,1) norm.
APPLIED INTELLIGENCE
(2022)
Article
Quantum Science & Technology
Davis Arthur, Prasanna Date
Summary: By formulating the training problem as a QUBO problem and utilizing adiabatic quantum computers, the quantum approach to solving balanced k-means clustering training problem shows promising characteristics, although the solutions obtained from quantum computers are generally not as good as those from classical algorithms. Scalability study and concept validation on the Iris benchmark dataset demonstrate the feasibility and potential of this quantum approach.
QUANTUM INFORMATION PROCESSING
(2021)
Article
Biology
Qinfen Wang, Geng Chen, Xuting Jin, Siyuan Ren, Gang Wang, Longbing Cao, Yong Xia
Summary: Mortality prediction is crucial in evaluating illness severity and improving patient prognosis. Existing methods for analyzing multivariate time series (MTSs) suffer from sparse and incomplete data. We propose a BiT-MAC network that captures both intra-time series coupling and inter-time series coupling to estimate missing values and improve MTS-based prediction. Extensive experiments on clinical datasets demonstrate the superiority of BiT-MAC and the interpretability of its features.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Engineering, Electrical & Electronic
Jiawei Li, Jinyuan Liu, Shihua Zhou, Qiang Zhang, Nikola K. Kasabov
Summary: In this paper, we propose a coordinated learning network for detail-refinement in multi-exposure image fusion. Our network obtains shallow feature maps from over/under-exposed source images and generates smooth attention weight maps to establish global connections. By cooperating with an edge revision module, our method effectively refines edge details and suppresses noise in fused images.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Ping Qiu, Yongshun Gong, Yuhai Zhao, Longbing Cao, Chengqi Zhang, Xiangjun Dong
Summary: This article explores an efficient method for mining negative sequential patterns (NSPs) using temporal point processes (TPPs) to model frequently occurring and nonoccurring events and behaviors. By loosening constraints, a new definition of negative containment is provided, and an efficient method for calculating the supports of negative sequences is proposed. Finally, a novel and efficient algorithm is presented to identify valuable NSPs.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Editorial Material
Computer Science, Artificial Intelligence
Longbing Cao
Summary: The uncertain world faces increasing emergencies, crises and disasters, including COVID-19 pandemic, hurricane Ian, global financial inflation and recession, misinformation disaster, and cyberattacks. AI for smart disaster resilience transforms traditional reactive and scripted disaster management into proactive and intelligent resilience in the face of diverse ECDs. This article provides a systematic overview of various ECDs, conventional ECD management, ECD data complexities, and the research landscape of AISDR. Translational disaster AI is crucial in enabling smart disaster resilience.
INTERNATIONAL JOURNAL OF DATA SCIENCE AND ANALYTICS
(2023)
Article
Computer Science, Artificial Intelligence
Longbing Cao
Summary: After years of development, a new generation of AI and data science has emerged, based on the integration of science, technology, and engineering. This new generation embraces Trans-AI/DS thinking, which combines AI and data science to promote transformative, transdisciplinary, and translational approaches. These paradigm shifts encourage innovative thinking beyond traditional AI and data-driven methods, and focus on the complexities of human intelligence, nature, society, and their creations.
INTERNATIONAL JOURNAL OF DATA SCIENCE AND ANALYTICS
(2023)
Article
Computer Science, Artificial Intelligence
Longbing Cao
Summary: After 70 years of AI and 50 years of DS, AI/DS have entered a new age, where they are built upon the integration of science, technology, and engineering. This integration has resulted in Trans-AI/DS, which promote transformative, transdisciplinary, and translational thinking, methodologies, and practices in AI/DS.
INTERNATIONAL JOURNAL OF DATA SCIENCE AND ANALYTICS
(2023)
Review
Biotechnology & Applied Microbiology
Grace Wen, Vickie Shim, Samantha Jane Holdsworth, Justin Fernandez, Miao Qiao, Nikola Kasabov, Alan Wang
Summary: This study explores the performance of various machine learning algorithms in harmonizing MRI data, summarizing the findings from relevant peer-reviewed articles. It provides guidelines for current methods and identifies potential future research directions. MRI data can be harmonized either implicitly (n = 21) or explicitly (n = 20).
BIOENGINEERING-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Sensen Song, Zhenhong Jia, Jie Yang, Nikola Kasabov
Summary: This paper proposes a fuzzy clustering method based on low-rank representation for image segmentation, which improves clustering results by enhancing superpixels constructed by edges and adding a fuzzy regularization term.
IEEE TRANSACTIONS ON FUZZY SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Jia Xu, Longbing Cao
Summary: This paper proposes a method that combines deep variational sequential learning with copula-based statistical dependence modeling to address the challenging problem of modeling high-dimensional, long-range dependencies between nonnormal multivariates. The method can characterize both the temporal dependence degrees and structures between the hidden variables representing the nonnormal multivariates, and it outperforms benchmarks in terms of both technical significance and portfolio forecasting performance.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Jiawei Li, Jinyuan Liu, Shihua Zhou, Qiang Zhang, Nikola K. Kasabov
Summary: Currently, multimodal medical image fusion technology has become an essential means for predicting diseases and studying pathology. To address the challenge of preserving unique features from different modal source images while ensuring time efficiency, a flexible semantic-guided architecture called GeSeNet is proposed. The experimental results demonstrate that our method outperforms ten state-of-the-art methods in generating high-quality fused images.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Zhilin Zhao, Longbing Cao, Kun-Yu Lin
Summary: In deep neural learning, training a discriminator on in-distribution samples may lead to misclassification of out-of-distribution samples, which poses a significant challenge for robust and safe deep learning. To address this issue, we propose a general approach called Fine-tuning Discriminators by Implicit Generators (FIG) that enhances the discriminatory power of standard discriminators in distinguishing in-distribution and out-of-distribution samples. FIG leverages information theory to infer an energy-based implicit generator from a discriminator and uses a Langevin dynamic sampler to draw specific out-of-distribution samples. Experimental results demonstrate that FIG achieves state-of-the-art out-of-distribution detection performance.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Zhilin Zhao, Longbing Cao, Kun-Yu Lin
Summary: Deep neural networks for image classification only learn to map in-distribution inputs to their corresponding ground-truth labels in training without differentiating out-of-distribution samples from in-distribution ones. To address this issue, we draw out-of-distribution samples from the vicinity distribution of training in-distribution samples for learning to reject the prediction on out-of-distribution inputs. Experiments show that the proposed method significantly outperforms existing methods in improving the capacity for discriminating between in-and out-of-distribution samples.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Jia Lei, Jiawei Li, Jinyuan Liu, Shihua Zhou, Qiang Zhang, Nikola K. Kasabov
Summary: The goal of multi-exposure image fusion is to generate synthetic results with abundant details and balanced exposure from low dynamic range (LDR) images. To solve the problem of existing methods only considering pixel values in local view field, we propose a global-local aggregation network for fusing extreme exposure images in an unsupervised way. Our method achieves the best results in terms of MEF-structure similarity index measure (SSIM) and peak signal-to-noise ratio (PSNR), outperforming 12 state-of-the-art fusion methods.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Computer Science, Artificial Intelligence
Zhilin Zhao, Longbing Cao, Chang-Dong Wang
Summary: The integrity of training data is uncertain, especially for non-IID datasets. Experts may misclassify samples, leading to unreliable labels. This study proposes a gray learning (GL) method that leverages both ground-truth and complementary labels to improve the robustness of neural networks.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Theory & Methods
Longbing Cao
Summary: This article provides an overview of the application of artificial intelligence techniques in the finance industry. It offers a comprehensive and dense landscape of the challenges, techniques, and opportunities of AIDS research in finance over the past decades. The article outlines the challenges of financial businesses and data, categorizes the decades of AIDS research in finance, illustrates the data-driven analytics and learning in financial businesses, compares classic and modern AIDS techniques, and discusses future opportunities for AIDS-empowered finance and finance-motivated AIDS research.
ACM COMPUTING SURVEYS
(2023)
Article
Computer Science, Artificial Intelligence
Rui Lv, Dingheng Wang, Jiangbin Zheng, Zhao-Xu Yang
Summary: In this paper, the authors investigate tensor decomposition for neural network compression. They analyze the convergence and precision of tensor mapping theory, validate the rationality of tensor mapping and its superiority over traditional tensor approximation based on the Lottery Ticket Hypothesis. They propose an efficient method called 3D-KCPNet to compress 3D convolutional neural networks using the Kronecker canonical polyadic (KCP) tensor decomposition. Experimental results show that 3D-KCPNet achieves higher accuracy compared to the original baseline model and the corresponding tensor approximation model.
Article
Computer Science, Artificial Intelligence
Xiangkun He, Zhongxu Hu, Haohan Yang, Chen Lv
Summary: In this paper, a novel constrained multi-objective reinforcement learning algorithm is proposed for personalized end-to-end robotic control with continuous actions. The approach trains a single model using constraint design and a comprehensive index to achieve optimal policies based on user-specified preferences.
Article
Computer Science, Artificial Intelligence
Zhijian Zhuo, Bilian Chen, Shenbao Yu, Langcai Cao
Summary: In this paper, a novel method called Expansion with Contraction Method for Overlapping Community Detection (ECOCD) is proposed, which utilizes non-negative matrix factorization to obtain disjoint communities and applies expansion and contraction processes to adjust the degree of overlap. ECOCD is applicable to various networks with different properties and achieves high-quality overlapping community detection.
Article
Computer Science, Artificial Intelligence
Yizhe Zhu, Chunhui Zhang, Jialin Gao, Xin Sun, Zihan Rui, Xi Zhou
Summary: In this work, the authors propose a Contrastive Spatio-Temporal Distilling (CSTD) approach to improve the detection of high-compressed deepfake videos. The approach leverages spatial-frequency cues and temporal-contrastive alignment to fully exploit spatiotemporal inconsistency information.
Review
Computer Science, Artificial Intelligence
Laijin Meng, Xinghao Jiang, Tanfeng Sun
Summary: This paper provides a review of coverless steganographic algorithms, including the development process, known contributions, and general issues in image and video algorithms. It also discusses the security of coverless steganography from theoretical analysis to actual investigation for the first time.
Article
Computer Science, Artificial Intelligence
Yajie Bao, Tianwei Xing, Xun Chen
Summary: Visual question answering requires processing multi-modal information and effective reasoning. Neural-symbolic learning is a promising method, but current approaches lack uncertainty handling and can only provide a single answer. To address this, we propose a confidence based neural-symbolic approach that evaluates NN inferences and conducts reasoning based on confidence.
Article
Computer Science, Artificial Intelligence
Anh H. Vo, Bao T. Nguyen
Summary: Interior style classification is an interesting problem with potential applications in both commercial and academic domains. This project proposes a method named ISC-DeIT, which combines data-efficient image transformer architectures and knowledge distillation, to address the interior style classification problem. Experimental results demonstrate a significant improvement in predictive accuracy compared to other state-of-the-art methods.
Article
Computer Science, Artificial Intelligence
Shashank Kotyan, Danilo Vasconcellos Vargas
Summary: This article introduces a novel augmentation technique called Dynamic Scanning Augmentation to improve the accuracy and robustness of Vision Transformer (ViT). The technique leverages dynamic input sequences to adaptively focus on different patches, resulting in significant changes in ViT's attention mechanism. Experimental results demonstrate that Dynamic Scanning Augmentation outperforms ViT in terms of both robustness to adversarial attacks and accuracy against natural images.
Article
Computer Science, Artificial Intelligence
Hiba Alqasir, Damien Muselet, Christophe Ducottet
Summary: The article proposes a solution to improve the learning process of a classification network by providing shape priors, reducing the need for annotated data. The solution is tested on cross-domain digit classification tasks and a video surveillance application.
Article
Computer Science, Artificial Intelligence
Dexiu Ma, Mei Liu, Mingsheng Shang
Summary: This paper proposes a method using neural dynamics solvers to solve infinity-norm optimization problems. Two improved solvers are constructed and their effectiveness and superiority are demonstrated through theoretical analysis and simulation experiments.
Article
Computer Science, Artificial Intelligence
Francesco Gregoretti, Giovanni Pezzulo, Domenico Maisto
Summary: Active Inference is a computational framework that uses probabilistic inference and variational free energy minimization to describe perception, planning, and action. cpp-AIF is a header-only C++ library that provides a powerful tool for implementing Active Inference for Partially Observable Markov Decision Processes through multi-core computing. It is cross-platform and improves performance, memory management, and usability compared to existing software.
Article
Computer Science, Artificial Intelligence
Zelin Ying, Dawei Cheng, Cen Chen, Xiang Li, Peng Zhu, Yifeng Luo, Yuqi Liang
Summary: This paper proposes a novel stock market trends prediction framework called SMART, which includes a self-supervised stock technical data sequence embedding model S3E. By training with multiple self-supervised auxiliary tasks, the model encodes stock technical data sequences into embeddings and uses the learned sequence embeddings for predicting stock market trends. Extensive experiments on China A-Shares market and NASDAQ market prove the high effectiveness of our model in stock market trends prediction, and its effectiveness is further validated in real-world applications in a leading financial service provider in China.
Article
Computer Science, Artificial Intelligence
Hao Li, Hao Jiang, Dongsheng Ye, Qiang Wang, Liang Du, Yuanyuan Zeng, Liu Yuan, Yingxue Wang, C. Chen
Summary: DHGAT1, a dynamic hyperbolic graph attention network, utilizes hyperbolic metric properties to embed dynamic graphs. It employs a spatiotemporal self-attention mechanism and weighted node representations, resulting in excellent performance in link prediction tasks.
Article
Computer Science, Artificial Intelligence
Jiehui Huang, Zhenchao Tang, Xuedong He, Jun Zhou, Defeng Zhou, Calvin Yu-Chian Chen
Summary: This study proposes a progressive learning multi-scale feature blending model for image deraining tasks. The model utilizes detail dilation and texture extraction to improve the restoration of rainy images. Experimental results show that the model achieves near state-of-the-art performance in rain removal tasks and exhibits better rain removal realism.
Article
Computer Science, Artificial Intelligence
Lizhi Liu, Zilin Gao, Yinhe Wang, Yongfu Li
Summary: This paper proposes a novel discrete-time interconnected model for depicting complex dynamical networks. The model consists of nodes and edges subsystems, which consider the dynamic characteristic of both nodes and edges. By designing control strategies and coupling modes, the stabilization and synchronization of the network are achieved. Simulation results demonstrate the effectiveness of the proposed methods.