检索结果-内蒙古大学图书馆

International Conference on Field Programmable Logic and Applications

作者： Chengming Zhang Tong Geng Anqi Guo Jiannan Tian Martin Herbordt Ang Lit Dingwen Tao School of Electrical Engineering and Computer Science Washington State University Pullman WA USA School of Informatics Computing and Engineering Indiana University Bloomington IN USA Mathematics and Computer Science Division Pacific Northwest National Laboratory Richland WA USA Department of Electrical and Computer Engineering University of Rochester Rochester NY USA Department of Electrical and Computer Engineering Boston University Boston MA USA

Graph Neural Networks (GNNs) have drawn tremendous attention due to their unique capability to extend Machine Learning (ML) approaches to applications broadly-defined as having unstructured data, especially graphs. Compared with other Machine Learning (ML) modalities, the acceleration of Graph Neural Networks (GNNs) is more challenging due to the irregularity and heterogeneity derived from graph typologies. Existing efforts, however, have focused mainly on handling graphs' irregularity and have not studied their heterogeneity. To this end we propose H-GCN, a PL (Programmable Logic) and AIE (AI Engine) based hybrid accelerator that leverages the emerging heterogeneity of Xilinx Versal Adaptive Compute Acceleration Platforms (ACAPs) to achieve high-performance GNN inference. In particular, H-GCN partitions each graph into three subgraphs based on its inherent heterogeneity, and processes them using PL and AIE, respectively. To further improve performance, we explore the sparsity support of AIE and develop an efficient density-aware method to automatically map tiles of sparse matrix-matrix multiplication (SpMM) onto the systolic tensor array. Compared with state-of-the-art GCN accelerators, H-GCN achieves, on average, speedups of 1.1~2.3x.

关键词： Tensors Adaptive systems Machine learning computer architecture Systolic arrays Graph neural networks Sparse matrices

来源：评论

学校读者我要写书评

暂无评论

Traffic Classification based on Incremental Learning Algorithms for the Software-Defined Networks

Traffic Classification based on Incremental Learning Algorit...

引用

2022 International Conference on Frontiers of Information Technology, FIT 2022

作者： Eldhai, Arwa M. Hamdan, Mosab Khan, Suleman Hamzah, Muzaffar Marsono, M.N. University of Science and Technology Faculty of Engineering Khartoum11115 Sudan Universiti Malaysia Sabah Faculty of Computing and Informatics Sabah Kota Kinabalu88450 Malaysia University of São Paulo Department of Computer Science São Paulo05508-090 Brazil University of Central Lancashire School of Psychology and Computer Science PrestonPR1 2HE United Kingdom Universiti Teknologi Malaysia Faculty of Electrical Engineering Johor81310 Malaysia

ISBN: (纸本)9798350345933

A new era of network administration has been ushered in by recent developments in software-defined networks (SDN) and traffic classification (TC) using machine learning (ML) techniques. All network devices are centrally managed and accessible through the SDN, which can simplify the TC process. The traditional data mining/ML approach that uses TC assumes that all of the task's data is always accessible and can be viewed simultaneously without processing time and memory restrictions. Therefore, these approaches are not effective in the case of stream learning since, in more realistic settings, the data is not available all at once and has a distinct distribution. Consequently, incremental learning algorithms (ILAs) can handle online data mining. This study's primary goal is to contrast various ILA approaches to enhance SDN's TC performance. In this study, we propose four ILAs: the self- adjusting memory coupled with the k Nearest Neighbor (kNN) classifier (SAMKNNC), the very fast decision rules classifier (VFDRC), the extremely fast decision tree classifier (EFDTC), and the streaming random patches ensemble classifier (SRPC). Both real and synthetic datasets are used for validation. Experimental findings reveal that the proposed techniques perform better in SDN traffic classification since they can effectively identify drift and use less memory and time. © 2022 IEEE.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Investigation of Industrial Bearing Fault Diagnosis Based on 1D-Cnn-Lstm

Investigation of Industrial Bearing Fault Diagnosis Based on...

引用

International Conference on Industrial engineering, Applications and Manufacturing (ICIEAM)

作者： Eduard Muratbakeev Diana Novak Yuriy Kozhubaev Department of Descriptive Geometry and Graphics Empress Catherine II Saint Petersburg Mining University Saint Petersburg Russia Department of Informatics and Computer Technologies Empress Catherine II Saint Petersburg Mining University Saint Petersburg Russia Department of Electrical Power Engineering and Electromechanics Empress Catherine II Saint Petersburg Mining University Saint Petersburg Russia

ISBN: (数字)9798331512194

ISBN: (纸本)9798331512200

This paper investigates a hybrid model method based on one-dimensional convolutional neural network (1DCNN) and long short-term memory (LSTM) network for fault diagnosis of industrial bearings. Based on a bearing dataset provided by Southeastern University, feature extraction and fault classification based on vibration signal data under different operating conditions were performed. The 1D-CNN module extracts local features from the signal in the time domain, while the LSTM module additionally captures the time dependence of the signal to improve the accuracy of fault diagnosis. Experimental results show that the model can effectively distinguish different types of bearing faults, exhibits good classification performance on the test set, and is highly robust to noise. This method provides an effective deep learning-based solution for industrial bearing fault diagnosis and has good application prospects.

关键词： Fault diagnosis Vibrations Training Accuracy Memory management Feature extraction Data models Stability analysis Convolutional neural networks Long short term memory

来源：评论

学校读者我要写书评

暂无评论

Streamlining Video Summarization with NLP: Techniques, Implementation, and Future Direction

引用

SN computer Science 2025年第2期6卷 1-17页

作者： Alqurni, Jehad Saad Alsmadi, Mutasem K. Alfagham, Hayat Alzoubi, Sharaf Ihab, Sohayla Sameh, Ahmed AbdElminaam, Diaa Salama Khalaf, Osamah Ibrahim Department of Educational Technologies College of Education Imam Abdulrahman Bin Faisal University P.O. Box 1982 Dammam Saudi Arabia Department of MIS College of Applied Studies and Community Service Imam Abdulrahman Bin Faisal University P.O. Box 1982 Dammam Saudi Arabia College of Computer Sciences and Informatics Amman Arab University Amman 11953 Jordan Faculty of Engineering Ain Shams University Cairo 19648 Egypt MEU Research Unit Middle East University Amman 11831 Lebanon Jadara Research Center Jadara University Irbid 21110 Jordan Department of Solar Al-Nahrain Research Center for Renewable Energy Al-Nahrain University Baghdad 42115 Iraq

The rapid growth of digital video content presents significant challenges in information accessibility and consumption, creating a pressing need for efficient video summarization techniques. This paper explores using natural language processing (NLP) to enhance video summarization by leveraging associated textual data such as subtitles, titles, and descriptions. Video summarization can be broadly divided into two main approaches: abstractive and extractive. We focus on extractive summarization, implementing various NLP techniques, including Pure NLTK, TextRank, LexRank, KL-Sum, and Naïve Reduction, each encapsulated in dedicated pipelines to tokenize, rank, and extract essential content. Subsequently, summaries are generated and presented as shortened video formats. Our methodology includes comparing these NLP-based techniques with current state-of-the-art approaches and evaluating them through various quantitative metrics such as Rouge, BLEU, and F1-score to ensure a comprehensive summarization quality and efficiency assessment. The study also addresses computational trade-offs, particularly focusing on optimizing summarization for real-time applications. Recognizing the reliance on textual data, we propose enhancements for handling videos with limited text, such as integrating audio-to-text conversion and visual analysis. Our findings underscore the potential of NLP-driven summarization in improving accessibility across varied video content types and provide a roadmap for future research to enhance scalability, personalization, and real-time applicability. The evaluation results demonstrate the effectiveness and potential of NLP techniques in video summarization regardless of video length, accent, or language. The findings highlight the effect of various NLP techniques on the form of the generated summaries. In addition, a comparison with state-of-the-art methodologies is performed to provide clearer insights into the quality of the summaries, not only regarding a

关键词： Abstractive summarization Audio-to-text conversion Computational efficiency Extractive summarization Multimedia content analysis Natural language processing (NLP) Personalization Real-time processing Scalability Video accessibility Video summarization

来源：评论

学校读者我要写书评

暂无评论

RSMA Inspired User Cooperation in Hybrid VLC/RF Networks for Coverage Extension

RSMA Inspired User Cooperation in Hybrid VLC/RF Networks for...

引用

IEEE Conference on Wireless Communications and Networking

作者： Konstantinos G. Rallis Vasilis K. Papanikolaou Sotiris A. Tegos Alexis A. Dowhuszko Panagiotis D. Diamantoulakis Mohammad-Ali Khalighi George K. Karagiannidis Department of Electrical and Computer Engineering Aristotle University of Thessaloniki Thessaloniki Greece Department of Applied Informatics University of Macedonia Thessaloniki Greece Department of Communications and Networking Aalto University Espoo Finland Aix-Marseille University CNRS Centrale Marseille Institut Fresnel Marseille France

In this paper, we propose and evaluate a hybrid visible light communication (VLC)/radio-frequency (RF) network architecture, where a VLC access point serves two user equipments (UEs), which also act as RF relays in order to extend the network’s coverage to a third UE outside the VLC cell. The proposed protocol is inspired by uplink rate-splitting multiple access to efficiently route the messages to the UEs. In more detail, the proposed protocol utilizes the nuances of the specific network architecture to efficiently utilize the wireless resource blocks for both coverage and throughput. The protocol is then optimized by maximizing the minimum achievable rate. Simulation results show that the proposed method achieves superior results compared with a more conventional benchmark scheme, that is also optimized under the same constraints.

关键词： Radio frequency Wireless communication Simulation Microprocessors computer architecture Network architecture Benchmark testing

来源：评论

学校读者我要写书评

暂无评论

Development of Mutal Recognition of Education Approach in Advanced computer-Oriented engineering Technologies in the Baltic Region

Development of Mutal Recognition of Education Approach in Ad...

引用

International Scientific Conference on Power and Electrical engineering of Riga Technical University (RTUCON)

作者： Nadezhda Kunicina Anton Rassõlkin Rasa Bruzgiene Darius Plonis Jelena Caiko Institute of Electrical Engineering and Electronics FEEE Riga Technical University Riga Latvia Department of Electrical Power Engineering and Mechatronics Tallinn University of Technology Tallinn Estonia Department of Computer Sciences Faculty of Informatics Kaunas University of Technology Kaunas Lithuania Department of Electronic Systems Vilnius Gediminas Technical University Vilnius Lithuania

The continued improvement and development of study programmes content for mutual recognition of study programmes among the largest technical universities in the Baltic region provides a very good opportunity for regional development as it brings a good advancement in academics and research at the involved universities, allows for more flexible and reliable collaboration with industrial partners, and provides an opportunity to develop new technologies and implement them for regional support. The goal of the development is to promote the involvement of students in the design of adaptive systems by providing them with the necessary set of skills that will allow them to immediately engage in the design of innovative adaptive systems and successfully solve technological challenges that have been identified through the cooperation of educational and research institutions. The adoption of an innovative educational strategy will aid in the creation of new technology and creative solutions even before students begin their studies. Collaboration between academic personnel and industry will help students build their abilities.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Using HLPC for parallelization of autonomous tests of WEB applications from its GUI 34

Using HLPC for parallelization of autonomous tests of WEB ap...

引用

34th European Modeling and Simulation Symposium, EMSS 2022

作者： Rossainz-López, Mario Islas-Fuentes, Jesús A. Pineda-Torres, Ivo Capel-Tuñón, Manuel Faculty of Computer Science Autonomous University of Puebla Av. San Claudio and 14 Sur Street San Manuel PueblaC.P. 72570 Mexico Software Engineering Department College of Informatics and Telecommunications ETSIIT University of Granada Daniel Saucedo Aranda s/n Granada18071 Spain

A proposal for parallelization of autonomous tests of Web applications from its graphical user interface (GUI) is presented to reduce its execution time since these increases exponentially due to the number of combinations that are generated with the different states of the fields, of the forms on the web application pages by creating a tree-like structure. It is proposed to use the model of high-level parallel compositions or HLPC as a suitable model of semi-automatic parallelization that defines a Treelike structure as a communication pattern between processes. The proposed HLPC, which we call HLPC-Tree, uses reinforcement learning that associates each node of the process tree as a slave object using the Q-Learning (QL) algorithm and achieves autonomous recognition of the fields of the forms and valid-invalid options to identify failures and display them with HTTP status codes. In addition, the Mechanize library is used to find the number of possible combinations through the states of the fields of the forms and to know how many nodes are generated at each level of the process tree in the HLPC-Tree when it grows in depth. Finally, the performance analysis of the proposed HLPC is shown with an analysis of the speedup and execution times in an 8-core machine to demonstrate good scalability in its accelerations compared to Amdahl's Law. © 2022 The Authors.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Compact Vertically Stacked Converter Module with Thermal Balancing and High-Power Dissipation Capability

A Compact Vertically Stacked Converter Module with Thermal B...

引用

Annual IEEE Conference on Applied Power Electronics Conference and Exposition (APEC)

作者： Jingyuan Liang Qi Liu Tiantian Liu Wai Tung Ng The Edward S. Rogers Sr. Department of Electrical & Computer Engineering University of Toronto Toronto Canada School of Integrated Circuits Southeast University Nanjing China State Key Laboratory of Functional Materials for Informatics Chinese Academy of Sciences Shanghai China

A compact vertically stacked buck converter is designed with improved thermal performance. Three copper core PCBs and two vertical double-diffused MOSFETs (VDMOS), in TO-Leaded Topside-Cooled (TOLT) package, are stacked together to form the half-bridge output stage. The inductor, the input and output capacitors, and the gate driver ICs are also assembled in this module. For comparison purposes, a horizontally laid out buck converter with the same components using a conventional FR4 PCB is also built. Under forced air cooling, the proposed vertically stacked 48 to 12 V buck converter, switching at 355 kHz, can achieve an output power of more than 160 W with a peak temperature of 81 degree Celsius. The same buck converter in horizontal configuration can only achieve a maximum output power of 60 W with a peak temperature of 90 degree Celsius. At 60 W output, the hottest point for the proposed vertically stacked converter is 23 degree Celsius lower than that of the conventional PCB layout. Moreover, the power conversion efficiency of the proposed vertically stacked converter is improved at the same 60 W power level. Switching at 244 kHz with optimized deadtimes, the maximum output power of the vertically stacked converter is measured to be 291 W. Benefited from the high-power dissipation capability, the converter can maintain a nearly constant efficiency across a wide range of output power from 194 to 291 W.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Predicting Career Trajectories of Information System Graduates Using AdaBoost Algorithm

Predicting Career Trajectories of Information System Graduat...

引用

International Conference on Wireless and Telematics (ICWT)

作者： Rahmad Kurniawan Siska Kurnia Gusti Sukamto Syamsudhuha Rahmad Abdillah Department of Computer Science Faculty of Mathematics and Natural Sciences Universitas Riau Pekanbaru Indonesia Department of Informatics Engineering Universitas Islam Negeri Sultan Syarif Kasim Riau Pekanbaru Indonesia Department of Mathematics Faculty of Mathematics and Natural Sciences Universitas Riau Pekanbaru Indonesia

The undergraduate unemployment rate in Indonesia is increasing, necessitating universities to produce quality graduates. In alignment with global technological advancements, universities are committed to enhancing the quality of their graduates through curriculum evaluation. In this study, we investigated the department of Information Systems at the University of Riau and developed an alum career analysis system using the AdaBoost algorithm. The data used in the study was collected via an online survey from graduates. Relevant features were selected from features to preprocess the data. The experiments conducted using the SAMME.R boosting method, with a learning rate of 0.1, 100 estimators, and cross-validation on the five-fold sampling method, resulted in the highest accuracy of 83.3%. The decision tree of the AdaBoost algorithm explains the factors influencing alums to pursue careers in their field, with English proficiency, learning methods, alum cohorts, and alum's activeness in job searching being the main determinants of success in their careers. The alum career analysis system proved reliable for evaluating the curriculum and producing quality graduates in line with technological advancements. Future work includes expanding the system to include graduates from other departments and implementing findings to improve the curriculum.

关键词：

来源：评论

学校读者我要写书评

暂无评论

In the Fog: Application Deployment for the Cloud Continuum 12

In the Fog: Application Deployment for the Cloud Continuum

引用

12th International Conference on Information, Intelligence, Systems and Applications, IISA 2021

作者： Apostolou, Dimitris Verginadis, Yiannis Mentzas, Gregoris University of Piraeus Deprtment of Informatics Piraeus Greece Athens University of Economics and Business Department of Business Administration Athens Greece National Technical University of Athens School of Electrical and Computer Engineering Athens Greece

ISBN: (纸本)9781665400329

Serverless and the Function-as-a-Service (FaaS) paradigms are seen as two enabling technologies for next-generation computing on the cloud continuum. This article discusses prominent frameworks to deploy and monitor applications that span the cloud continuum. It discusses associated challenges and proposes a novel architecture for a framework that manages intelligently multi-cloud, fog and edge resources in order to cope with the requirements of FaaS-enabled applications and services. © 2021 IEEE.

关键词： Fog computing

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：