检索结果-内蒙古大学图书馆

Accelerated Fuzzy C-Means Clustering Based on New Affinity Filtering and Membership Scaling

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Li, Dong Zhou, Shuisheng Pedrycz, Witold School of Mathematics and Statistics Xidian University Xi’an710071 China The Department of Electrical and Computer Engineering University of Alberta EdmontonT6R 2V4 AB Canada The Systems Research Institute Polish Academy of Sciences Warsaw00-901 Poland The Department of Electrical and Computer Engineering Faculty of Engineering King Abdulaziz University Jeddah21589 Saudi Arabia

Fuzzy C-Means (FCM) is a widely used clustering method. However, FCM and its many accelerated variants have low efficiency in the mid-to-late stage of the clustering process. In this stage, all samples are involved in the update of their non-affinity centers, and the fuzzy membership grades of the most of samples, whose assignment is unchanged, are still updated by calculating the samples-centers distances. All those lead to the algorithms converging slowly. In this paper, a new affinity filtering technique is developed to recognize a complete set of the non-affinity centers for each sample with low computations. Then, a new membership scaling technique is suggested to set the membership grades between each sample and its non-affinity centers to 0 and maintain the fuzzy membership grades for others. By integrating those two techniques, FCM based on new affinity filtering and membership scaling (AMFCM) is proposed to accelerate the whole convergence process of FCM. Many experimental results performed on synthetic and real-world data sets have shown the feasibility and efficiency of the proposed algorithm. Compared with the state-of-the-art algorithms, AMFCM is significantly faster and more effective. For example, AMFCM reduces the number of the iteration of FCM by 80% on average. Copyright © 2023, The Authors. All rights reserved.

关键词： Efficiency

Challenges and Opportunities of Machine Learning on Neutron and X-ray Scattering

学校读者我要写书评

暂无评论

Synchrotron Radiation News 2022年第4期35卷 16-20页

作者： Drucker, Nathan C. Liu, Tongtong Chen, Zhantao Okabe, Ryotaro Chotrattanapituk, Abhijatmedhi Nguyen, Thanh Wang, Yao Li, Mingda Quantum Measurement Group MIT Cambridge MA United States School of Engineering and Applied Sciences Harvard University Cambridge MA United States Department of Physics MIT Cambridge MA United States Department of Mechanical Engineering MIT Cambridge MA United States Department of Chemistry MIT Cambridge MA United States Department of Electrical Engineering and Computer Science MIT Cambridge MA United States Department of Nuclear Science and Engineering MIT Cambridge MA United States Department of Physics and Astronomy Clemson University Clemson SC United States

Deep Internal Learning: Deep Learning from a Single Input

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Tirer, Tom Giryes, Raja Chun, Se Young Eldar, Yonina C. Faculty of Engineering Bar-Ilan University Ramat Gan5290002 Israel School of Electrical Engineering Tel Aviv University Tel Aviv69978 Israel Department of Electrical and Computer Engineering Seoul National University 1 Gwanak-ro Gwanak-gu Seoul08826 Korea Republic of Faculty of Mathematics and Computer Science Weizmann Institute of Science Rehovot7610001 Israel

Deep learning, in general, focuses on training a neural network from large labeled datasets. Yet, in many cases there is value in training a network just from the input at hand. This is particularly relevant in many signal and image processing problems where training data is scarce and diversity is large on the one hand, and on the other, there is a lot of structure in the data that can be exploited. Using this information is the key to deep internal-learning strategies, which may involve training a network from scratch using a single input or adapting an already trained network to a provided input example at inference time. This survey paper aims at covering deep internal-learning techniques that have been proposed in the past few years for these two important directions. While our main focus will be on image processing problems, most of the approaches that we survey are derived for general signals (vectors with recurring patterns that can be distinguished from noise) and are therefore applicable to other modalities. Copyright © 2023, The Authors. All rights reserved.

关键词： Large datasets

Noise-Resilient Designs for Optical Neural Networks

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Kosmella, Gianluca Stabile, Ripalta Sanders, Jaron Department of Electrical Engineering Eindhoven University of Technology PO Box 513 Eindhoven5600 MB Netherlands Department of Mathematics and Computer Science Eindhoven University of Technology PO Box 513 Eindhoven5600 MB Netherlands

All analog signal processing is fundamentally subject to noise, and this is also the case in modern implementations of Optical Neural Networks (ONNs). Therefore, to mitigate noise in ONNs, we propose two designs that are constructed from a given, possibly trained, Neural Network (NN) that one wishes to implement. Both designs have the capability that the resulting ONNs gives outputs close to the desired NN. To establish the latter, we analyze the designs mathematically. Specifically, we investigate a probabilistic framework for the first design that establishes that the design is correct, i.e., for any feed-forward NN with Lipschitz continuous activation functions, an ONN can be constructed that produces output arbitrarily close to the original. ONNs constructed with the first design thus also inherit the universal approximation property of NNs. For the second design, we restrict the analysis to NNs with linear activation functions and characterize the ONNs’ output distribution using exact formulas. Finally, we report on numerical experiments with LeNet ONNs that give insight into the number of components required in these designs for certain accuracy gains. We specifically study the effect of noise as a function of the depth of an ONN. The results indicate that in practice, adding just a few components in the manner of the first or the second design can already be expected to increase the accuracy of ONNs *** Codes 68T07 © 2023, CC BY.

关键词： Chemical activation

Solution formula for generalized two-phase Stokes equations and its applications to maximal regularity;model problems

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Kajiwara, Naoto Applied Physics Course Department of Electrical Electronic and Computer Engineering Gifu University Yanagido 1-1 Gifu501-1193 Japan

In this paper we give a solution formula for the two-phase Stokes equations with and without surface tension and gravity in the whole space with flat interface. The solution formula has already considered by Shibata-Shimizu. However we reconstruct the formula so that we are able to prove resolvent estimate and maximal regularity estimate. In the previous work, they needed to assume additional conditions on normal components. We also take care of normal components, while the assumption becomes weaker than before. The method is based on an H∞-calculus which has already used for the Stokes problems with various boundary conditions in the half *** Codes 35D35, 35B45 © 2022, CC0.

关键词： Navier Stokes equations

Teleoperation of Multi-Degrees-of-Freedom Manipulator through Composite State Convergence Scheme

学校读者我要写书评

暂无评论

IFAC-PapersOnLine 2022年第1期55卷 126-130页

作者： Muhammad Usman Asad Jason Gu Umar Farooq Rajeeb Dey Valentina E. Balas Ghulam Abbas Department of Electrical and Computer Engineering Dalhousie University Halifax NS B3H 4R2 Canada Electrical Engineering National Institute of Technology Silchar Assam India Department of Automatics and Applied Software “Aurel Vlaicu” University of Arad Arad Romania Department of Electrical Engineering The University of Lahore Lahore Pakistan

In this paper, design of composite state convergence controller for teleoperating a multi-degrees-of-freedom manipulator is presented. The composite variables corresponding to the joints of the master and slave manipulators as well as operator’s force are transmitted across the communication channel. The control laws for the master and slave systems are defined according to the composite state convergence scheme with the addition of cancellation terms containing the nonlinear dynamics of master and slave manipulators. It is shown that convergence of composite variables of respective master and slave joints guarantee the convergence of respective joint positions of the master and slave systems under the composite state convergence control laws. MATLAB simulations on time-delayed teleoperation model are performed to validate the proposal.

关键词： Teleoperation state convergence composite variables time delay stability

Negative Binomial Regression Model for Predicting the Number of Tourists in National Parks of Thailand 10

学校读者我要写书评

暂无评论

Negative Binomial Regression Model for Predicting the Number...

10th International Conference on Digital Arts, Media and Technology and 8th ECTI Northern Section Conference on electrical, Electronics, computer and Telecommunications engineering, ECTI DAMT and NCON 2025

作者： Thepsumritporn, Sthaporn Jitpattanakul, Anuchit Mekruksavanich, Sakorn Phaphan, Wikanda King Mongkut's University of Technology North Bangkok Faculty of Applied Science Department of Applied Statistics Bangkok10800 Thailand King Mongkut's University of Technology North Bangkok Faculty of Applied Science Department of Mathematics Bangkok10800 Thailand Intelligent and Nonlinear Dynamic Innovations Research Center Thailand University of Phayao School of Information and Communication Department of Computer Engineering Phayao56000 Thailand Research Group in Statistical Learning and Inference Thailand

ISBN: (纸本)9798331543273

This article examines the use of count data models to predict the number of tourists visiting Thailand's national parks. The dataset, encompassing observations from 2016 to 2022 across 146 national parks, exhibits characteristics such as discreteness, over dispersion, and non-negative integer values. Two models - Poisson (POI), and Negative Binomial (NB) - were compared to identify the most effective approach for modeling and prediction. Model evaluation metrics, including the Akaike Information Criterion (AIC) and log-likelihood, were employed. Results demonstrate that the Negative Binomial regression model best captures the data's overdispersion compared to the POI model. The NB model achieved the lowest AIC and highest log-likelihood, affirming its superior performance for this application. Our findings underscore the NB model's utility in predicting tourism-related count data and suggest its broader applicability to other domains characterized by overdispersed count data. © 2025 IEEE.

关键词： Logistic regression

Two-Dimensional Thin Layer Convective Flow and Heat Transfer of a Dusty Fluid on a Nonlinear Stretching Sheet

学校读者我要写书评

暂无评论

Multiscale Science and engineering 2022年第3期4卷 111-118页

作者： Islam, Md. Robiul Ferdows, M. Misra, J. C. Murtaza, M. G. Research Group of Fluid Flow Modeling and Simulation Department of Applied Mathematics University of Dhaka Dhaka Bangladesh Department of Computer Science and Engineering Green University of Bangladesh Dhaka Bangladesh Centre for Healthcare Science and Technology Indian Institute of Engineering Science and Technology Howrah India Department of Mathematics Comilla University Cumilla Bangladesh

In this study, steady two-dimensional boundary layer flow and heat transfer of a dusty fluid over a nonlinearly stretching sheet with variable thickness, $${u}_{w} (x)={U}_{0}{(x+b)}^{m}$$ has been considered. A variable magnetic field is applied transversely to the direction of the flow. The general problem has been reduced to a system of coupled non-linear ordinary differential equations of two-dimensional boundary layer. By means of similarity transformation, the basic governing equations are transformed into a set of nonlinear ordinary differential equation and then are solved numerically by dsolve routine in MAPLE. Results of the present study have been duly validated by adopting an appropriate procedure. The influence of various parameters such as velocity power index, wall thickness parameter, magnetic parameter fluid-particle interaction parameter for velocity, fluid-particle interaction parameter for temperature, mass concentration of dust particle, Prandtl number, Eckert number has been investigated. The effects are then presented graphically both for dust phase and fluid phase. It is clear that flow field and other quantities of physical interest are significantly influenced by these parameters. Also the effects of the parameters on the skin-friction coefficients and rate of heat transfer in terms of the Nusselt number, which are of physical interest, are shown graphically/in tabular form. The study reveals that the impact of the parameters for the dusty fluid on Nusselt number is considerably higher than that for a clean fluid.

关键词：

Maximal Lp-Lq regularity for the Stokes equations with various boundary conditions in the half space

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Kajiwara, Naoto Applied Physics Course Department of Electrical Electronic and Computer Engineering Gifu University Yanagido 1-1 Gifu501-1193 Japan

We prove resolvent Lp estimates and maximal Lp-Lq regularity estimates for the Stokes equations with Dirichlet, Neumann and Robin boundary conditions in the half space. Each solution is constructed by a Fourier multiplier of x′-direction and an integral of xN-direction. We decompose the solution such that the symbols of the Fourier multipliers are bounded and holomorphic. We see that the operator norms are dominated by a homogeneous function of order −1 for xN-direction. The basis are Weis's operator-valued Fourier multiplier theorem and a boundedness of a kernel operator. We give a new simple approach to get maximal regularity in the half *** Codes 35D35, 35B45 © 2022, CC0.

关键词： Navier Stokes equations