检索结果-内蒙古大学图书馆

MedFILIP: Medical Fine-grained Language-Image Pre-training

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Liang, Xinjie Li, Xiangyu Li, Fanding Jiang, Jie Dong, Qing Wang, Wei Wang, Kuanquan Dong, Suyu Luo, Gongning Li, Shuo School of Computer Science and Technology Harbin Institute of Technology Harbin China Department of Thoracic Surgery at No. 4 Affiliated Hospital Harbin Medical University Harbin China School of Computer Science and Technology Harbin Institute of Technology Shenzhen China College of computer and control engineering Northeast Forestry University Harbin China Computer Electrical and Mathematical Sciences & Engineering Division King Abdullah University of Science and Technology Thuwal Saudi Arabia Department of Biomedical Engineering Department of Computer and Data Science Case Western Reserve University ClevelandOH United States

Medical vision-language pretraining (VLP) that leverages naturally-paired medical image-report data is crucial for medical image analysis. However, existing methods struggle to accurately characterize associations between images and diseases, leading to inaccurate or incomplete diagnostic results. In this work, we propose MedFILIP, a fine-grained VLP model, introduces medical image-specific knowledge through contrastive learning, specifically: 1) An information extractor based on a large language model is proposed to decouple comprehensive disease details from reports, which excels in extracting disease deals through flexible prompt engineering, thereby effectively reducing text complexity while retaining rich information at a tiny cost. 2) A knowledge injector is proposed to construct relationships between categories and visual attributes, which help the model to make judgments based on image features, and fosters knowledge extrapolation to unfamiliar disease categories. 3) A semantic similarity matrix based on fine-grained annotations is proposed, providing smoother, information-richer labels, thus allowing fine-grained image-text alignment. 4) We validate MedFILIP on numerous datasets, e.g., RSNA-Pneumonia, NIH ChestX-ray14, VinBigData, and COVID-19. For single-label, multi-label, and fine-grained classification, our model achieves state-of-the-art performance, the classification accuracy has increased by a maximum of 6.69%. The code is available in https://***/ PerceptionComputingLab/MedFILIP. Copyright © 2025, The Authors. All rights reserved.

关键词： Semantics

Electro osmotically flow of fourth-grade fluid in complex channel with porous medium and lubricated walls: applications in petroleum industry

学校读者我要写书评

暂无评论

Waves in Random and Complex Media 2023年

作者： Nazeer, Mubbashar Pasha, Amjad Ali Hussain, Farooq Adil, Mohammad Mechanical Engineering Program Physical Science and Engineering Division King Abdullah University of Science and Technology Thuwal Saudi Arabia BUITEMS Quetta Pakistan KAUST Clean Combustion Research Center King Abdullah University of Science and Technology Thuwal Saudi Arabia Department of Mathematics Institute of Arts and Sciences Government College University Faisalabad Chiniot Campus Pakistan Aerospace Engineering Department King Abdulaziz University Jeddah Saudi Arabia

This article offers a distinct survey of a multiphase flow through a uniform channel. The electro-osmotic phenomenon is the main source of the flow. Fourth-grade fluid model is used as the base liquid which is suspended with metallic particles to form a bi-phase viscous suspension. Skin friction effects are tackled by employing lubrication effects on the opposite walls of the channel. Porous media is governed by the modified Darcy’s law and the Poisson differential equation is considered for electric double-layer effects. An approximate solution is achieved for the nonlinear fluid flow. The obtained findings are also compared with the existing literature for the limiting case. It is inferred that the shear thickness effects cause the momentum of the bi-phase flow to decline, gradually. However, the motion of the metallic particles is supported by the lubrication effects on the walls of the channel. © 2023 Informa UK Limited, trading as Taylor & Francis Group.

关键词： Porous materials

Expressivity and Approximation Properties of Deep Neural Networks with ReLUk Activation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： He, Juncai Mao, Tong Xu, Jinchao Computer Electrical and Mathematical Science and Engineering Division King Abdullah University of Science and Technology Thuwal23955 Saudi Arabia Department of Mathematics The Pennsylvania State University University ParkPA16802 United States

In this paper, we investigate the expressivity and approximation properties of deep neural networks employing the ReLUk activation function for k ≥ 2. Although deep ReLU networks can approximate polynomials effectively, deep ReLUk networks have the capability to represent higher-degree polynomials precisely. Our initial contribution is a comprehensive constructive proof for polynomial representation using deep ReLUk networks. This allows us to establish an upper bound on both the size and count of network parameters. Consequently, we are able to demonstrate a suboptimal approximation rate for functions from Sobolev spaces as well as for analytic functions. Additionally, through an exploration of the representation power of deep ReLUk networks for shallow networks, we reveal that deep ReLUk networks can approximate functions from a range of variation spaces, extending beyond those generated solely by the ReLUk activation function. This finding demonstrates the adaptability of deep ReLUk networks in approximating functions within various variation spaces. © 2023, CC BY.

关键词： Deep neural networks

An Efficient Automated Technique for Classification of Breast Cancer Using Deep Ensemble Model

学校读者我要写书评

暂无评论

computer Systems science & engineering 2023年第7期46卷 897-911页

作者： Muhammad Zia Ur Rehman Jawad Ahmad Emad Sami Jaha abdullah Marish Ali Mohammed A.Alzain Faisal Saeed Department of Electrical Engineering HITEC UniversityTaxila47080Pakistan School of Computing Edinburgh Napier UniversityEdinburghEH105DTUK Department of Computer Science Faculty of Computing and Information TechnologyKing Abdulaziz UniversityJeddah21589Saudi Arabia Department of Information Technology College of Computers and Information TechnologyTaif UniversityTaif21944Saudi Arabia DAAI Research Group Department of Computing and Data ScienceSchool of Computing and Digital TechnologyBirmingham City University15 Bartholomew RowBirminghamB55JUUK

Breast cancer is one of the leading cancers among *** has the second-highest mortality rate in women after lung *** detection,especially in the early stages,can help increase survival ***,manual diagnosis of breast cancer is a tedious and time-consuming process,and the accuracy of detection is reliant on the quality of the images and the radiologist’s ***,computer-aided medical diagnosis has recently shown promising results,leading to the need to develop an efficient system that can aid radiologists in diagnosing breast cancer in its early *** research presented in this paper is focused on the multi-class classification of breast *** deep transfer learning approach has been utilized to train the deep learning models,and a pre-processing technique has been used to improve the quality of the ultrasound *** proposed technique utilizes two deep learning models,Mobile-NetV2 and DenseNet201,for the composition of the deep ensemble *** learning models are fine-tuned along with hyperparameter tuning to achieve better ***,entropy-based feature selection is *** cancer identification using the proposed classification approach was found to attain an accuracy of 97.04%,while the sensitivity and F1 score were 96.87%and 96.76%,*** performance of the proposed model is very effective and outperforms other state-of-the-art techniques presented in the literature.

关键词： Breast cancer image enhancement ensemble model transfer learning feature selection

VERY HIGH-ORDER A-STABLE STIFFLY ACCURATE DIAGONALLY IMPLICIT RUNGE-KUTTA METHODS WITH ERROR ESTIMATORS

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Alamri, Yousef Ketcheson, David I. Computer Electrical and Mathematical Sciences & Engineering Division King Abdullah University of Science and Technology Thuwal23955 Saudi Arabia

A numerical search approach is used to design high-order diagonally implicit Runge-Kutta (DIRK) schemes equipped with embedded error estimators, some of which have identical diagonal elements (SDIRK) and explicit first stage (ESDIRK). In each of these classes, we present new A-stable schemes of order six (the highest order of previously known A-stable DIRK-type schemes) up to order eight. For each order, we include one scheme that is only A-stable as well as schemes that are L-stable, stiffly accurate, and/or have stage order two. The latter types require more stages, but give better convergence rates for differential-algebraic equations (DAEs), and those which have stage order two give better accuracy for moderately stiff problems. The development of the eighth-order schemes requires, in addition to imposing A-stability, finding highly accurate numerical solutions for a system of 200 equations in over 100 variables, which is accomplished via a combination of global and local optimization strategies. The accuracy, stability, and adaptive stepsize control of the schemes are demonstrated on diverse problems. Copyright © 2022, The Authors. All rights reserved.

关键词： Runge Kutta methods

Analysis and Assessment of Wind Energy Potential of Al-Hodeidah in Yemen

学校读者我要写书评

暂无评论

computers, Materials & Continua 2021年第11期69卷 1995-2011页

作者： Fahd N.Al-Wesabi Murad A.Almekhlafi Mohammed abdullah Al-Hagery Mohammad Alamgeer Khalid Mahmood Majdy M.Eltahir Ali M.Al-Sharafi Amin M.El-Kustaban Department of Computer Science King Khalid UniversityMuhayel AseerKingdom of Saudi Arabia Faculty of Computer and IT Sana’a UniversitySana’aYemen Department of Electrical Engineering–Communication Engineering Sana’a UniversityYemen Department of Computer Science College of ComputerQassim UniversitySaudi Arabia Department of Information Systems King Khalid UniversityMayahel AseerKSA Department of Computer Science College of Computers and Information TechnologyUniversity of BishaKSA Department of Electronic Faculty of EngineeringUSTSana’aYemen

Renewable energy is one of the essential elements of the social and economic development in any civilized *** use of fossil fuels and the non-renewable form of energy has many adverse effects on the most of *** the high potential of renewable energy sources in Yemen and the absence of similar studies in the region,this study aimed to examine the wind energy potential of Hodeidah-Yemen Republic by analyzing wind characteristics and assessment,determining the available power density,and calculate the wind energy extracted at different *** average wind speed of Hodeidah was obtained only for the data currently available for the five years 2005–2009(due to the current economic and the political situation in Yemen).The results show that the average wind speed in the five years is(25.2 W/m2 at 10 m,93.9 W/m2 at 30 m,and 173.5 W/m2 at 50 m).The average yearly wind power density(25.2 W/m2 at 10 m,93.9 W/m2 at 30 m and 173.5 W/m2 at 50m),and the average yearly energy density(220.8 KWh/m2/year at 10 m,822.6 KWh/m2/year at 30 m and 1519.9 KWh/m2/year at 50 m).This research is a preliminary assessment of the potential of wind energy in Hodeidah,which provides useful information for developing wind energy and an efficient wind *** to the International Wind Energy Rating criteria,the region of Hodeidah falls under‘Class 2’and is classified as‘Marginal’for most of the year.

关键词： Energy potential wind speed weibull distribution wind power density

Thermal simulation of microwave kiln based on multiphysics

学校读者我要写书评

暂无评论

Thermal simulation of microwave kiln based on multiphysics

2023 International Technical Conference on Circuits/Systems, computers, and Communications, ITC-CSCC 2023

作者： Jansaengsuk, Thodsaphon Pattanapichai, Sorathon Poopanya, Piyawong Phimphakan, Nonthawat Thongsri, Jatuporn College of Advanced Manufacturing Innovation King Mongkut's Institute of Technology Ladkrabang Computer Simulation in Engineering Research Group Bangkok10520 Thailand Ubon Ratchathani Rajabhat University Program of Physics Faculty of Science Ubonratchatani34000 Thailand

ISBN: (纸本)9798350326413

A microwave kiln, made of silicon carbide and ceramic fiber, commonly employs in a household glassware production process. In this process, when the kiln was in a microwave oven, a microwave transmitted to the kiln generating a high temperature to fuse the glass inside. This article presents a thermal simulation to investigate the temperature inside the kiln based on multiphysics consisting of a high-frequency structure simulator (HFSS) and computational fluid dynamics (CFD). The multiphysics results revealed the temperature inside the kiln in a transient state, consistent with the experimental results. As expected, the temperature increased with the increasing time of the process. Significantly, the silicon carbide had higher temperatures than the ceramic fiber;therefore, silicon carbide is a crucial material for generating heat inside the kiln. In addition, the temperature-increasing rate (TIR) inside the kiln depended on the kiln's thickness (Th). As a result, the thinner Th provided better TIR. The research findings can be applied to develop a high-efficacy household glassware production process. © 2023 IEEE.

关键词： Multiphysics

SD-Transformer: A System-Level Denoising Transformer for Encrypted Traffic Behavior Identification

学校读者我要写书评

暂无评论

SD-Transformer: A System-Level Denoising Transformer for Enc...

IEEE Conference on Global Communications (GLOBECOM)

作者： Yizhuo Zhao Yukun Zhu Xiong Li Ruidong Chen Mohammad S. Obaidat Pandi Vijayakumar School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China King Abdullah II School of Information Technology University of Jordan Amman Jordan Department of Computer Science and Engineering University College of Engineering Tindivanam Tindivanam India

Encrypted behavior identification is crucial in ensuring network security. Most existing solutions in this area recognize behavior by observing encrypted traffic patterns between users and applications. However, such solutions rely on features such as timing, packet sequence, and packet length, which may be affected by network fluctuations, and thus have weak generalization capabilities. In this paper, we first analyze the impact of noise on the network, such as parameters and network delays during API requests. By combining a noise-based traffic collector with an improved Transformer model, we propose a system-level denoising Transformer method for encrypted traffic behavior identification called SD-Transformer. It is able to filter system noise by utilizing an attention mechanism and targeted noise packet masking. We evaluate the performance of SD-Transformer on three datasets, i.e., ISCX-VPN, USTC-TFC, and our generated noise-containing Web Application Traffic dataset (WEB-APP), and it achieves an accuracy of 95.97%, 93.59%, and 99.82%, respectively. Besides, compared to the state-of-the-art methods, the accuracy is increased to 96.82% (↑16.0%) and 85.41% (↑17.76%) on the WEB-APP dataset under different API parameters and network latency environments, respectively. Additionally, the target mask of the SD-Transformer achieves 96.45% accuracy with an improvement of 11.29% on the WEB-APP dataset with latency.

关键词：

Deep Neural Networks and Finite Elements of Any Order on Arbitrary Dimensions

学校读者我要写书评

暂无评论

arXiv 2023年

作者： He, Juncai Xu, Jinchao Computer Electrical and Mathematical Science and Engineering Division King Abdullah University of Science and Technology Thuwal23955 Saudi Arabia Department of Mathematics The Pennsylvania State University University ParkPA16802 United States

In this study, we establish that deep neural networks employing ReLU and ReLU2 activation functions can effectively represent Lagrange finite element functions of any order on various simplicial meshes in arbitrary dimensions. We introduce two novel formulations for globally expressing the basis functions of Lagrange elements, tailored for both specific and arbitrary meshes. These formulations are based on a geometric decomposition of the elements, incorporating several insightful and essential properties of high-dimensional simplicial meshes, barycentric coordinate functions, and global basis functions of linear elements. This representation theory facilitates a natural approximation result for such deep neural networks. Our findings present the first demonstration of how deep neural networks can systematically generate general continuous piecewise polynomial functions on both specific or arbitrary simplicial *** Codes 68T07, 65D40 Copyright © 2023, The Authors. All rights reserved.

关键词： Mesh generation