Article
Mathematics, Applied
Braxton Osting, Dong Wang, Yiming Xu, Dominique Zosso
Summary: Archetypal analysis is an unsupervised learning method that utilizes convex polytopes to summarize multivariate data, with archetype points being the key components. Consistency results are proven, showing convergence of archetype points under certain assumptions, along with convergence rates for optimal objective values. Experiments with various distributions support the analysis and demonstrate the effectiveness of the method for summarizing data.
SIAM JOURNAL ON MATHEMATICS OF DATA SCIENCE
(2021)
Article
Automation & Control Systems
Abdul Suleman
Summary: The study compared three approaches to archetypal analysis and found that the original method is generally more accurate in uncovering cluster structures in data, but this superiority is not clear for data generated from real-life problems and dedicated to unsupervised clustering problems.
INTERNATIONAL JOURNAL OF FUZZY SYSTEMS
(2021)
Article
Multidisciplinary Sciences
Mikio Shiga, Shigeto Seno, Makoto Onizuka, Hideo Matsuda
Summary: Single-cell RNA sequencing technology allows us to understand biological processes at unprecedented resolution, but it requires a complex data processing pipeline. Unsupervised cell clustering plays a crucial role in identifying cell types, discovering cell diversity, and subpopulations. The use of different quantification methods to process gene expression profiles can significantly impact clustering results.
Article
Computer Science, Artificial Intelligence
Wenhui Wu, Junhui Hou, Shiqi Wang, Sam Kwong, Yu Zhou
Summary: In this paper, we propose a semi-supervised adaptive kernel concept factorization (SAKCF) method that integrates data representation and kernel learning and solves the problem using an alternating iterative algorithm. Experimental results demonstrate the effectiveness and advantages of SAKCF over other methods in clustering tasks.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Artificial Intelligence
Joseph A. Gallego, Fabio A. Gonzalez, Olfa Nasraoui
Summary: This paper explores the correspondence between least-square estimation in a reproducing kernel Hilbert space and different M-estimators in the original space, as well as the application of new robust kernels associated with various types of M-estimators in clustering tasks. The results show that some robust kernels perform as well as state-of-the-art robust clustering methods.
Article
Automation & Control Systems
Zhiwei Xing, Meng Wen, Jigen Peng, Jinqian Feng
Summary: The paper introduces a novel discriminative semi-supervised NMF (DSSNMF) algorithm that effectively utilizes label information from a portion of the data, with empirical experiments demonstrating its effectiveness.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2021)
Article
Computer Science, Information Systems
Siyuan Peng, Zhijing Yang, Bingo Wing-Kuen Ling, Badong Chen, Zhiping Lin
Summary: A new semi-supervised NMF method called dual semi-supervised convex nonnegative matrix factorization (DCNMF) is proposed in this paper. DCNMF incorporates the pointwise and pairwise constraints of labeled samples into convex NMF, resulting in a better low-dimensional data representation. It can process mixed-sign data due to the nonnegative constraint only on the coefficient matrix.
INFORMATION SCIENCES
(2022)
Article
Engineering, Electrical & Electronic
Zhenwen Ren, Mithun Mukherjee, Mehdi Bennis, Jaime Lloret
Summary: In this paper, a novel non-negative matrix factorization tailored graph tensor MKGC method (TMKGC) is proposed to address the challenge of clustering non-linear data more effectively. TMKGC integrates NMF and graph learning in kernel space to learn multiple candidate affinity graphs and captures the high-order structure information of all candidate graphs in a 3-order tensor kernel space using tensor singular value decomposition based tensor nuclear norm. Extensive experiments demonstrate the superiority of TMKGC over existing MKGC methods.
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Ping Deng, Fan Zhang, Tianrui Li, Hongjun Wang, Shi-Jinn Horng
Summary: Clustering remains a challenging research hotspot in data mining. This paper proposes a biased unconstrained non-negative matrix factorization (BUNMF) model to improve the clustering performance. The model modifies the update rules and adds bias, and introduces three different activation functions for iteration updates.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Khanh Luong, Richi Nayak, Thirunavukarasu Balasubramaniam, Md Abul Bashar
Summary: This paper proposes a deep non-negative matrix factorization-based framework for effective multi-view data clustering by uncovering the non-linear relationships and intrinsic components of the data. The framework effectively incorporates the optimal manifold of multi-view data and outperforms existing multi-view matrix factorization-based methods.
PATTERN RECOGNITION
(2022)
Article
Computer Science, Artificial Intelligence
Ping Deng, Tianrui Li, Dexian Wang, Hongjun Wang, Hong Peng, Shi-Jinn Horng
Summary: Multi-view clustering based on non-negative matrix factorization (NMFMvC) is a well-known method for handling high-dimensional multi-view data. However, the optimization method using the Karush-Kuhn-Tucker (KKT) conditions is poorly scalable. In this study, we propose an unconstrained non-negative matrix factorization multi-view clustering (uNMFMvC) model, which decouples the elements of the matrix and combines them with a non-linear mapping function in a non-negative value domain. The objective function is optimized using the stochastic gradient descent (SGD) algorithm, and three uNMFMvC methods are constructed based on different mapping functions.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Lin Feng, Wenzhe Liu, Xiangzhu Meng, Yong Zhang
Summary: This paper introduces an innovative multi-view clustering method, SMCTN, which utilizes triplex regularized non-negative matrix factorization to effectively extract multi-view information while maintaining low-dimensional geometry structure. Extensive experimental results on textual and image datasets demonstrate the superior performance of the proposed method.
Article
Computer Science, Information Systems
Dexian Wang, Tianrui Li, Ping Deng, Fan Zhang, Wei Huang, Pengfei Zhang, Jia Liu
Summary: In this article, a Generalized Deep Learning Clustering (GDLC) algorithm based on Non-Negative Matrix Factorization (NMF) is proposed to address the slow convergence and low clustering accuracy in the update process of NMF. A nonlinear constrained NMF (NNMF) algorithm is constructed to achieve sequential updates of the matrix elements guided by the learning rate, and the GDLC algorithm is constructed by transforming gradient values into generalized weights and biases through a nonlinear activation function. The experimental results on eight datasets demonstrate the efficient performance of the GDLC algorithm.
ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA
(2023)
Article
Engineering, Electrical & Electronic
Jicong Fan, Chengrun Yang, Madeleine Udell
Summary: The paper introduces a new robust nonlinear factorization method called Robust Non-Linear Matrix Factorization (RNLMF), which is capable of denoising and clustering tasks, with robustness to sparse noise and outliers, and demonstrates significant improvements over baseline methods in empirical experiments.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Mathematics, Applied
Li Chen, Shuisheng Zhou, Jiajun Ma, Mingliang Xu
Summary: Kernel-based clustering algorithms have the advantage of identifying and capturing nonlinear structures in datasets, but face challenges in handling large-scale datasets due to memory constraints. In this study, an incomplete Cholesky factorization method is proposed to generate low-rank approximations of kernel matrices, accelerating kernel clustering and saving memory space. Experimental results show that the proposed algorithm achieves similar performance to traditional kernel k-means clustering, but with the capability to handle large-scale datasets.
APPLIED MATHEMATICS AND COMPUTATION
(2021)
Article
Behavioral Sciences
Christian S. Musaeus, Knut Engedal, Peter Hogh, Vesna Jelic, Arjun R. Khanna, Troels Wesenberg Kjaer, Morten Morup, Mala Naik, Anne-Rita Oeksengaard, Emiliano Santarnecchi, Jon Snaedal, Lars-Olof Wahlund, Gunhild Waldemar, Birgitte B. Andersen
BRAIN AND BEHAVIOR
(2020)
Article
Mathematics, Interdisciplinary Applications
Kristoffer Jon Albers, Morten Morup, Mikkel N. Schmidt, Fumiko Kano Gluckstad
Summary: Data-driven segmentation is a crucial tool for analyzing patterns of associations in social survey data, and its quality can be quantified by its ability to predict held-out data. Comparing different methods, we found that data-driven segmentation outperforms demographic markers in predicting human values segmentation, and Bayesian Latent Class Analysis (LCA) performs better than the standard maximum likelihood LCA and is more robust for different numbers of clusters.
JOURNAL OF MATHEMATICAL SOCIOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Greta Tuckute, Sofie Therese Hansen, Troels Wesenberg Kjaer, Lars Kai Hansen
Summary: This study implemented a neurofeedback training paradigm during a sustained visual attention task, using real-time scalp EEG signals to decode attentional states. The neurofeedback group showed higher levels of task-relevant attentional information in the brain before making correct behavioral responses compared to incorrect responses. A portable EEG neurofeedback system was developed to decode attentional states and predict behavioral choices, which is open source and allows for active engagement in further development of neurofeedback tools.
NEURAL COMPUTATION
(2021)
Article
Neurosciences
Kristoffer J. Albers, Karen S. Ambrosen, Matthew G. Liptrot, Tim B. Dyrby, Mikkel N. Schmidt, Morten Morup
Summary: This study proposes a framework for quantitative evaluation of brain parcellations via statistical prediction of connectomics data, revealing substantial differences in parcellation structures that characterize functional and structural connectivity. The choice of fine-grained and coarse representations used by existing atlases is crucial, with resolution being more critical than exact border location of parcels.
Article
Computer Science, Artificial Intelligence
Petr Taborsky, Laurent Vermue, Maciej Korzepa, Morten Morup
Summary: The article introduces a novel Bayesian probabilistic model for graph cutting, providing an effective solution to separating community structures in complex networks. The method demonstrates excellent performance on real social networks and image segmentation problems, while also learning the parameter space.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2021)
Article
Neurosciences
Kristoffer Jon Albers, Matthew G. Liptrot, Karen Sando Ambrosen, Rasmus Roge, Tue Herlau, Kasper Winther Andersen, Hartwig R. Siebner, Lars Kai Hansen, Tim B. Dyrby, Kristoffer H. Madsen, Mikkel N. Schmidt, Morten Morup
Summary: Through multi-modal integration, a consensus representation can be obtained that well explains both functional and structural connectomes, providing improved representations of functional connectivity compared to using functional data alone.
FRONTIERS IN NEUROSCIENCE
(2022)
Article
Biochemical Research Methods
Nicolai F. Pedersen, Torsten Dau, Lars Kai Hansen, Jens Hjortkjaer
Summary: Temporal synchrony between facial motion and acoustic modulations is a key characteristic of audiovisual speech. This study used regularized canonical correlation analysis to investigate the precise rates at which envelope information is synchronized with motion in different parts of the face. The results revealed distinct bandpass speech envelope filters at different temporal scales, with one set correlated with mouth movements and another set correlated with global face and head motion.
PLOS COMPUTATIONAL BIOLOGY
(2022)
Article
Neurosciences
Anders S. Olsen, Rasmus M. T. Hoegh, Jesper L. Hinrich, Kristoffer H. Madsen, Morten Morup
Summary: The article presents a multimodal, multisubject directional archetypal analysis method for modeling metastable microstates in EEG/MEG data. The method extends the traditional archetypal analysis by modeling the continuous trajectories of microstates and accounting for scale and polarity invariance.
FRONTIERS IN NEUROSCIENCE
(2022)
Article
Neurosciences
Anders S. Olsen, Anders Lykkebo-Valloe, Brice Ozenne, Martin K. Madsen, Dea S. Stenbaek, Sophia Armand, Morten Morup, Melanie Ganz, Gitte M. Knudsen, Patrick M. Fisher
Summary: This study evaluated the impact of psilocin on the characteristics of resting-state time-varying functional connectivity in healthy individuals. The findings suggest that specific brain states showing negative associations with drug level and subjective drug intensity contribute to a better understanding of the acute effects of serotonergic psychedelics.
Editorial Material
Neurosciences
Camillo Porcaro, Kamran Avanaki, Oscar Arias-Carrion, Morten Morup
FRONTIERS IN NEUROSCIENCE
(2023)
Proceedings Paper
Acoustics
Anders S. Olsen, Emil Ortvald, Kristoffer H. Madsen, Mikkel N. Schmidt, Morten Morup
Summary: The development of suitable models for dynamic functional connectivity is crucial for a better understanding of the brain's activity during rest and tasks. This study introduces mixture models and Hidden Markov models that consider the sign-symmetric distribution of eigenvectors on a hypersphere, and demonstrates their performance on synthetic and task-fMRI data.
2023 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING WORKSHOPS, ICASSPW
(2023)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Nikolaos Nakis, Abdulkadir Celikkanat, Morten Morup
Summary: A central aim of modeling complex networks is to accurately embed networks in order to detect structures and predict link and node properties. The Latent Space Model (LSM) has become a prominent framework for embedding networks and includes the Latent Distance Model (LDM) and Eigenmodel (LEM) as the most widely used LSM specifications. We present the Hybrid-Membership Latent Distance Model (HM-LDM) as a reconciliation of LSMs with latent community detection, demonstrating its effectiveness in accurate node representations and community extraction.
COMPLEX NETWORKS AND THEIR APPLICATIONS XI, COMPLEX NETWORKS 2022, VOL 1
(2023)
Proceedings Paper
Engineering, Biomedical
Ali Mohebbi, Anna-Katharina Boehm, Jens Magelund Tarp, Morten Lind Jensen, Henrik Bengtsson, Morten Morup
Summary: The study aims to explore the recovery of consensus metrics using less than 14 days of CGM data, finding relatively low deviations for time in range and average based metrics with less than 14 days, but large deviations in metrics characterizing infrequent events. Additionally, clear discrepancies were observed in consensus metrics obtained in two consecutive 14 day periods.
2021 43RD ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE & BIOLOGY SOCIETY (EMBC)
(2021)
Article
Social Sciences, Mathematical Methods
Mikkel N. Schmidt, Daniel Seddig, Eldad Davidov, Morten Morup, Kristoffer Jon Albers, Jan Michael Bauer, Fumiko Kano Gluckstad
Summary: This study examined the optimal number of clusters for characterizing distinctive Schwartz value typologies using LPA, finding that eight clusters generate meaningful insights and predict external variables.
METHODOLOGY-EUROPEAN JOURNAL OF RESEARCH METHODS FOR THE BEHAVIORAL AND SOCIAL SCIENCES
(2021)
Meeting Abstract
Endocrinology & Metabolism
A. Mohebbi, H. Bengtsson, M. Jensen, B. Stallknecht, N. -K. Kjoller, M. Morup
DIABETES TECHNOLOGY & THERAPEUTICS
(2021)
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.