Article
Computer Science, Artificial Intelligence
Jingzheng Li, Hailong Sun
Summary: Unsupervised domain adaptation methods aim to enhance feature transferability but may sacrifice feature discriminability. This study proposes Noise-robust cross-domain Contrastive Learning (NaCL) to simultaneously learn instance-wise discrimination and encoding semantic structures for domain adaptation task.
Article
Computer Science, Artificial Intelligence
Hongfu Liu, Junxiang Chen, Jennifer Dy, Yun Fu
Summary: K-means is a widely used clustering algorithm known for its simplicity and efficiency. This review paper focuses on generalizing K-means to solve challenging and complex problems. It unifies the available approaches in terms of data representation, distance measure, label assignment, and centroid updating. Concrete applications of modified K-means formulations are reviewed, including iterative subspace projection and clustering, consensus clustering, constrained clustering, domain adaptation, and outlier detection.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Engineering, Electrical & Electronic
Yihong Cao, Hui Zhang, Xiao Lu, Yurong Chen, Zheng Xiao, Yaonan Wang
Summary: Unsupervised domain adaptation is an effective approach for solving the labeling difficulties in semantic segmentation tasks. A novel clustering-based method is proposed, which uses an adaptive refining-aggregation-separation framework to learn discriminative features for different domains and features. This method does not require tunable thresholds and includes techniques such as adaptive refinement, feature evaluation, and different losses for improving segmentation performance. Experimental results on benchmark datasets show that the proposed method outperforms existing state-of-the-art methods.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Engineering, Multidisciplinary
Bernardo P. Ferreira, F. M. Andrade Pires, M. A. Bessa
Summary: This article introduces adaptivity in Clustering-based Reduced Order Models (ACROMs) and applies it to Self-Consistent Clustering Analysis (SCA). The Adaptive Self-Consistent Clustering Analysis (ASCA) method improves predictions for materials with history-dependent localization phenomena. The method consists of three main building blocks and proposes solutions to further enhance the adaptive process.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2022)
Article
Physics, Multidisciplinary
G. Tucci, E. Roldan, A. Gambassi, R. Belousov, F. Berger, R. G. Alonso, A. J. Hudspeth
Summary: Modeling noisy oscillations of active systems is a challenge in physics and biology. A linear stochastic model driven by non-Markovian bistable noise is proposed and shown to generate self-sustained periodic oscillation. Experimental data on hair bundles in bullfrog sacculus support this minimal model accurately describing bistable-like oscillatory motion.
PHYSICAL REVIEW LETTERS
(2022)
Article
Multidisciplinary Sciences
Jay Vornhagen, Emily K. Roberts, Lavinia Unverdorben, Sophia Mason, Alieysa Patel, Ryan Crawford, Caitlyn L. Holmes, Yuang Sun, Alexandra Teodorescu, Evan S. Snitkin, Lili Zhao, Patricia J. Simner, Pranita D. Tamma, Krishna Rao, Keith S. Kaye, Michael A. Bachman
Summary: This study identifies several genes reproducibly associated with progression to infection in patients colonized by diverse Klebsiella. Patient variables, such as comorbidities, partially explain which patients will progress to Klebsiella infection, with colonization of the gut acting as a reservoir. Little is known, however, regarding Klebsiella genes that may increase risk of disease in colonized individuals.
NATURE COMMUNICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Wenwen Qiang, Jiangmeng Li, Changwen Zheng, Bing Su
Summary: This paper proposes a novel method called AT-MCAN for unsupervised domain adaptation, which introduces a covariance-aware divergence metric and an auxiliary clustering task to enhance the discriminability of features, allowing the classifier to utilize data from both domains effectively.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Information Systems
Ang Li, Shengsheng Wang, Xin Zhao, Juan Chen
Summary: Previous methods in unsupervised domain adaptation for semantic segmentation fail to consider the intra-domain gap among the target domain itself. This paper proposes a style clustering-based unsupervised domain adaptation method, which effectively captures the latent distributions within the target data and reduces the intra-domain gap. Experimental results on benchmark datasets demonstrate the effectiveness of the proposed method.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Gianni Costa, Riccardo Ortale
Summary: An innovative unsupervised approach is presented for the interrelationship of topic modeling with document clustering, utilizing Bayesian generative modeling and posterior inference. The approach seamlessly unifies and jointly carries out the two tasks, enabling automatic inference of the relationships between word-embedding topics and cluster components. Extensive empirical study on benchmark real-world corpora demonstrates the method's higher effectiveness in partitioning text collections and discovering their semantics.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Multidisciplinary Sciences
Maelle Jospin, Benjamin Bonneau, Viviane Laine, Jean-Louis Bessereau
Summary: The properties of Caenorhabditis elegans levamisole-sensitive acetylcholine receptors (L-AChRs) can be modified by their clustering machinery, which is important for the receptors' functionality in their native environment.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2022)
Article
Mathematics, Applied
Puneet Pasricha, Xiaoping Lu, Song-Ping Zhu
Summary: Ma and Xu (2016) proposed a Hawkes jump-diffusion model for the firm's value to describe the unexpectedness of default and default clustering in the framework of Merton's structural default. They presented a closed-form solution for the probability of default and the default correlation using the characteristic function, which can substantially improve computational efficiency for the problem.
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS
(2021)
Article
Agriculture, Dairy & Animal Science
Y. Steyn, T. Lawlor, Y. Masuda, S. Tsuruta, A. Legarra, D. Lourenco, I. Misztal
Summary: Maintaining genetic variation in a population is important for long-term genetic gain. The existence of subpopulations within a breed helps maintain genetic variation and diversity. Stratifying selected candidates into sub-populations using K-means clustering successfully separated genetically different groups.
JOURNAL OF DAIRY SCIENCE
(2023)
Article
Computer Science, Artificial Intelligence
Xiaobo Zhang, Tao Wang, Xiaole Zhao, Dengmin Wen, Donghai Zhai
Summary: In this study, a multitask-guided deep clustering (DC) with boundary adaptation (MTDC-BA) based on a convolutional neural network autoencoder (CNN AE) is proposed. The algorithm utilizes deep feature extraction and model parameter storage to capture multitask knowledge. It also incorporates boundary adaptation techniques and data augmentation to improve clustering performance. Experimental results demonstrate the superiority of MTDC-BA over traditional clustering methods in terms of both performance and efficiency.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Youwei Liang, Dong Huang, Chang-Dong Wang, Philip S. Yu
Summary: This article introduces a new framework for multi-view graph learning, which models the consistency and inconsistency of multiple views in a unified objective function. It effectively handles low-quality or noisy datasets by designing an efficient optimization algorithm that can obtain an approximate solution in linear time complexity. Experimental results demonstrate the robustness and efficiency of the proposed approach.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Thermodynamics
Sangjo Kim
Summary: A new performance adaptation method for aero gas turbine engines is proposed in this study to improve prediction accuracy by effectively processing a large amount of measured data. The adaptation factors are used to adjust compressor performance, bleed air flow, engine thrust, and exhaust gas temperature, and it is confirmed that this method can generate accurate gas turbine engine models using time series measurement data.