检索结果-内蒙古大学图书馆

IEEE/ACM Workshop on Memory Centric High Performance Computing (MCHPC)

作者： Benjamin Sepanski Tuowen Zhao Hans Johansen Samuel Williams Computer Science Department University of Texas at Austin Computer Science Department Univeristy of Utah Computational Research Division Lawrence Berkeley National Laboratory

ISBN: (纸本)9781665472838

Computations on structured grids using standard multidimensional array layouts can incur substantial data movement costs through the memory hierarchy. This paper explores the benefits of using a framework (Bricks) to separate the complexity of data layout and optimized communication from the functional representation. To that end, we provide three novel contributions and evaluate them on several kernels taken from GENE, a phase-space fusion tokamak simulation code. We extend Bricks to support 6-dimensional arrays and kernels that operate on complex data types, and integrate Bricks with cuFFT. We demonstrate how to optimize Bricks for data reuse, spatial locality, and GPU hardware utilization achieving up to a 2.67 × speedup on a single A100 GPU. We conclude with insights on how to rearchitect memory subsystems.

关键词： Costs High performance computing Layout graphics processing units Tokamak devices Spatial databases Hardware

来源：评论

学校读者我要写书评

暂无评论

Polar, Spherical and Orthogonal Space Subdivisions for an Algorithm Acceleration: O(1) Point-in-Polygon/Polyhedron Test

arXiv

引用

arXiv 2022年

作者： Skala, Vaclav The University of West Bohemia Faculty of Applied Sciences Department .of Computer Science and Engineering Center of Computer Graphics and Visualization Univerzitni 8 PlzenCZ 30614 Czech Republic

Acceleration of algorithms is becoming a crucial problem, if larger data sets are to be processed. Evaluation of algorithms is mostly done by using computational geometry approach and evaluation of computational complexity. However in today’s engineering problems this approach does not respect that number of processed items is always limited and a significant role plays also speed of read/write operations. One general method how to speed up an algorithm is application of space subdivision technique and usually the orthogonal space subdivision is used. In this paper non-orthogonal subdivisions are described. The proposed approach can significantly improve memory consumption and run-time complexity. The proposed modified space subdivision techniques are demonstrated on two simple problems Point-in-Convex Polygon and Point-in-Convex Polyhedron *** Codes 65D17, 65D18 © 2022, CC BY.

关键词： Computational geometry

来源：评论

学校读者我要写书评

暂无评论

Computation of an Emptiable Minimal Siphon in a Subclass of Petri Nets Using Mixed-Integer Programming

引用

IEEE/CAA Journal of Automatica Sinica 2021年第1期8卷 219-226页

作者： Shouguang Wang Wenli Duo Xin Guo Xiaoning Jiang Dan You Kamel Barkaoui MengChu Zhou the School of Information and Electronic Engineering Zhejiang Gongshang UniversityHangzhou 310018China the Macao Institute of Systems Engineering Macao University of Science and TechnologyMacao 999078China the Department of Computer Science Conservatoire National des Arts et MétiersParis 75141France the Department of Electrical and Computer Engineering New Jersey Institute of TechnologyNewarkNJ 07102 USA the Institute of Systems Engineering and Collaborative Laboratory for Intelligent Science and Systems Macao University of Science and TechnologyMacao 999078China

Deadlock resolution strategies based on siphon control are widely *** computational efficiency largely depends on siphon ***-integer programming(MIP)can be utilized for the computation of an emptiable siphon in a Petri net(PN).Based on it,deadlock resolution strategies can be designed without requiring complete siphon enumeration that has exponential *** to this reason,various MIP methods are proposed for various subclasses of *** work proposes an innovative MIP method to compute an emptiable minimal siphon(EMS)for a subclass of PNs named S^(4)*** particular,many particular structural characteristics of EMS in S4 PR are formalized as constraints,which greatly reduces the solution *** results show that the proposed MIP method has higher computational ***,the proposed method allows one to determine the liveness of an ordinary S^(4)PR.

关键词： Automated manufacturing systems deadlocks discrete event system mixed-integer programming(MIP) Petri nets(PN) siphons

来源：评论

学校读者我要写书评

暂无评论

Orthogonal projection for anomaly detection in networking datasets

引用

Journal of Ambient Intelligence and Humanized Computing 2023年第6期14卷 7957-7966页

作者： Cortes-Polo, David Jimenez, Luis I. Paoletti, Mercedes E. Calle-Cancho, Jesus Rico-Gallego, Juan A. Department of Signal Theory and Communication University Rey Juan Carlos Cam. del Molino 5 Madrid Fuenlabrada28942 Spain Intelligent Systems Group Department of Computer Science University of Valladolid P.O de Belén 15 Valladolid47011 Spain Hyperspectral Computing Laboratory Department of Technology of Computers and Communication University of Extremadura Av. de la Universidad s/n Caceres10003 Spain Calle Sola 1 Extremadura Trujillo10200 Spain Computer and Telematics Systems Engineering University of Extremadura Av. de la Universidad s/n Caceres10003 Spain

In recent years, the impressive growth of new wireless technologies, together with the appearance of new requirements in applications and services, is progressively changing the use of networks. Due to the high mobility required, the network must adapt to the infrastructure to meet the demands of the users. As a result, service providers currently have to over-provision network capacity, which is costly. In addition, considering efficient resource planning in advance involves a lot of labor-intensive efforts. Consequently, network usage analysis is a very useful tool that allows network administrators to find patterns and anomalies. Whilst pattern detection provides administrators the ability to define the infrastructure, anomaly detection provides rich and valuable information for certain applications, for example, to avoid network saturation in urban areas during peak hours. This article proposes a new methodology based on orthogonal projections over Call Detail Records (CDR) for anomaly detection to help in the dynamic management of the network in an urban area. The method is evaluated in a real scenario provided by an Italian telecommunications operator, considering different locations in the Milan metropolitan area, differentiated by the geographic resolution of the data, reaching F1 scores above 0.8. In addition, a new ground truth is presented, hoping it will become a reference data set for the community, in the form of a set of locations that have been corroborated for use in evaluating anomaly detection techniques. © 2023, The Author(s).

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Multi-class Classification of Motor Imagery EEG Signals Using Deep Learning Models

引用

SN computer science 2024年第5期5卷 444页

作者： Khemakhem, Rafik Belgacem, Sana Echtioui, Amira Ghorbel, Mohamed Ben Hamida, Ahmed Kammoun, Ines Advanced Technologies for Medicine and Signals Laboratory ‘ATMS’ National Engineering School of Sfax (ENIS) Sfax University Sfax Tunisia Higher Institute of Management of Gabès Gabès University Gabès Tunisia Department IS College of Computer Science King Khaled University ‘KKU’ Abha Saudi Arabia Functional Exploration Department of Habib Bourguiba Hospital Sfax University Sfax Tunisia

The accurate classification of Motor Imagery (MI) electroencephalography (EEG) signals is crucial for advancing Brain-computer Interface (BCI) technologies, particularly for individuals with disabilities. In this study, we present a sophisticated deep learning methodology that systematically evaluates three models CNN, RNN, and BiLSTM, to identify the optimal approach for MI signal classification. Leveraging the BCI Competition IV 2a dataset, we applied a pre-processing step removing three EOG channels and retaining 22 EEG channels, extracting 288 MI epochs, each lasting 3 s. Our findings highlight the superior performance of the proposed RNN model, achieving a remarkable maximum accuracy of 98%. This outcome signifies a significant advancement in MI signal classification, demonstrating the potential of deep learning techniques to enhance BCI precision. The study contributes by introducing a novel methodology and showcasing its efficacy through rigorous evaluation against benchmarks, providing valuable insights for the development of more robust and accurate BCI systems. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： BCI BiLSTM CNN EEG Motor imagery RNN

来源：评论

学校读者我要写书评

暂无评论

NUC-Net: Non-uniform Cylindrical Partition Network for Efficient LiDAR Semantic Segmentation

引用

IEEE Transactions on Circuits and Systems for Video Technology 2025年

作者： Wang, Xuzhi Feng, Wei Kong, Lingdong Wan, Liang Tianjin Normal University College of Computer and Information Engineering Tianjin300380 China Tianjin Key Laboratory of Student Mental Health and Intelligence Assessment China Tianjin University College of Intelligence and Computing Tianjin300350 China National University of Singapore Department of Computer Science 117597 Singapore

LiDAR semantic segmentation plays a vital role in autonomous driving. Existing voxel-based methods for LiDAR semantic segmentation apply uniform partition to the 3D Li- DAR point cloud to form a structured representation based on cartesian/cylindrical coordinates. Although these methods show impressive performance, the drawback of existing voxel-based methods remains in two aspects: (1) it requires a large enough input voxel resolution, which brings a large amount of computation cost and memory consumption. (2) it does not well handle the unbalanced point distribution of LiDAR point cloud. In this paper, we propose a non-uniform cylindrical partition network named NUC-Net to tackle the above challenges. Specifically, we propose the Arithmetic Progression of Interval (API) method to non-uniformly partition the radial axis and generate the voxel representation which is representative and efficient. Moreover, we propose a non-uniform multi-scale aggregation method to improve contextual information. Our method achieves state-of-the-art performance on SemanticKITTI and nuScenes datasets with much faster speed and much less training time. And our method can be a general component for LiDAR semantic segmentation, which significantly improves both the accuracy and efficiency of the uniform counterpart by 4× training faster and 2× GPU memory reduction and 3× inference speedup. We further provide theoretical analysis towards understanding why NUC is effective and how point distribution affects performance. Code is available at https://***/alanWXZ/NUC-Net. © 1991-2012 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

A RAG-BASED MULTI-AGENT LLM SYSTEM FOR NATURAL HAZARD RESILIENCE AND ADAPTATION

arXiv

引用

arXiv 2024年

作者： Xie, Yangxinyu Jiang, Bowen Mallick, Tanwi Bergerson, Joshua David Hutchison, John K. Verner, Duane R. Branham, Jordan Alexander, M. Ross Ross, Robert B. Feng, Yan Levy, Leslie-Anne Su, Weijie J. Taylor, Camillo J. Department of Statistics and Data Science University of Pennsylvania PhiladelphiaPA United States Department of Computer and Information Science University of Pennsylvania PhiladelphiaPA United States Mathematics and Computer Science Division Argonne National Laboratory LemontIL United States Decision and Infrastructure Sciences Division Argonne National Laboratory LemontIL United States Environmental Science Division Argonne National Laboratory LemontIL United States

Large language models (LLMs) are a transformational capability at the frontier of artificial intelligence and machine learning that can support decision-makers in addressing pressing societal challenges such as extreme natural hazard events. As generalized models, LLMs often struggle to provide context-specific information, particularly in areas requiring specialized knowledge. In this work, we propose a Retrieval-Augmented Generation (RAG)-based multi-agent LLM system to support analysis and decision-making in the context of natural hazards and extreme weather events. As a proof of concept, we present WildfireGPT, a specialized system focused on wildfire scenarios. The architecture employs a user-centered, multi-agent design to deliver tailored risk insights across diverse stakeholder groups. By integrating domain-specific projection data, observational datasets, and scientific literature through a RAG framework, the system ensures both accuracy and contextual relevance of the information it provides. Evaluation across ten expert-led case studies demonstrates that WildfireGPT significantly outperforms existing LLM-based solutions for decision support in natural hazard and extreme weather contexts. © 2024, CC BY-NC-SA.

关键词： Inference engines

来源：评论

学校读者我要写书评

暂无评论

An Enhanced Levenberg-Marquardt Method via Gram Reduction 39

An Enhanced Levenberg-Marquardt Method via Gram Reduction

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Liu, Chengchang Luo, Luo Lui, John C.S. Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong School of Data Science Fudan University China Shanghai Key Laboratory for Contemporary Applied Mathematics China

ISBN: (纸本)157735897X

This paper studies the problem of solving the system of nonlinear equations. We propose the Gram-reduced Levenberg-Marquardt method, which reuses the Gram matrix. Our method has a global convergence guarantee without relying on any step of line-search or solving sub-problems. We show that our method takes a smaller computational complexity than existing Levenberg-Marquardt methods to find the stationary point of the square norm of the equations. We also show that the proposed method enjoys a local superlinear convergence rate under the non-degenerate assumption. Experiments are conducted on real-world applications in scientific computing and machine learning, which validate the efficiency of our method. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： e-science

来源：评论

学校读者我要写书评

暂无评论

LGMD-based Neural Network for Detecting Abnormal Velocity Targets in Moving Crowd

LGMD-based Neural Network for Detecting Abnormal Velocity Ta...

引用

computer and Information Processing Technology (ISCIPT), International Symposium on

作者： Zhongxiang Zhao Bin Hu State Key Laboratory of Public Big Data College of Computer Science and Technology Guizhou University Guiyang P.R. China Department of Computer Science College of Computer Science & Technology Guizhou University Guiyang P.R. China Artificial Intelligence Research Institute of Guizhou University Guizhou University Guiyang P.R. China

ISBN: (数字)9798350388404

ISBN: (纸本)9798350388411

Detecting abnormal velocity targets is critical especially in public surveillance for crowd activity monitoring. Although there are some explorations on one such issue, traditional methods still perform poorly in complex scenes. In this paper, a novel visual neural network is investigated to perceive abnormal velocity targets in moving crowd, based on the latest neurophysiological achievements revealed in locusts’ vision systems. The proposed neural network contains two neural counterparts, i.e., the presynaptic and the postsynaptic networks. The former one receives visual signals and processes them to capture the motion cues of different targets, and the latter one filters the salience energies to perceive abnormal velocity targets in the field of view. Numerical experiments carried out show that the proposed neural network can effectively detect abnormal velocity targets in moving crowd. This study is an important step towards dynamic visual information processing in crowd behavior analysis.

关键词： Visualization computer vision Filters Machine vision Surveillance Neural networks Dynamics

来源：评论

学校读者我要写书评

暂无评论

IMPLICIT NEURAL REPRESENTATIONS FOR SIMULTANEOUS REDUCTION AND CONTINUOUS RECONSTRUCTION OF MULTI-ALTITUDE CLIMATE DATA

arXiv

引用

arXiv 2024年

作者： Qayyum, Alif Bin Abdul Luo, Xihaier Urban, Nathan M. Qian, Xiaoning Yoon, Byung-Jun Department of Electrical and Computer Engineering Texas A&M University United States Computational Science Initiative Brookhaven National Laboratory United States

The world is moving towards clean and renewable energy sources, such as wind energy, in an attempt to reduce greenhouse gas emissions that contribute to global warming. To enhance the analysis and storage of wind data, we introduce a deep learning framework designed to simultaneously enable effective dimensionality reduction and continuous representation of multi-altitude wind data from discrete observations. The framework consists of three key components: dimensionality reduction, cross-modal prediction, and super-resolution. We aim to: (1) improve data resolution across diverse climatic conditions to recover high-resolution details;(2) reduce data dimensionality for more efficient storage of large climate datasets;and (3) enable cross-prediction between wind data measured at different heights. Comprehensive testing confirms that our approach surpasses existing methods in both super-resolution quality and compression efficiency. © 2024, CC BY.

关键词： Dimensionality reduction

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：