检索结果-内蒙古大学图书馆

Skyway:Accelerate Graph Applications with a Dual-Path Architecture and Fine-Grained Data Management

Journal of computer science & technology 2024年第4期39卷 871-894页

作者： Mo Zou Ming-Zhe Zhang Ru-Jia Wang Xian-He Sun Xiao-Chun Ye Dong-Rui Fan Zhi-Min Tang Institute of Computing Technology Chinese Academy of SciencesBeijing 100190China University of Chinese Academy of Sciences Beijing 100049China Institute of Information Engineering Chinese Academy of SciencesBeijing 100045China Department of Computer Science Illinois Institute of TechnologyChicagoIL 60616U.S.A.

Graph processing is a vital component of many AI and big data ***,due to its poor locality and complex data access patterns,graph processing is also a known performance killer of AI and big data *** this work,we propose to enhance graph processing applications by leveraging fine-grained memory access patterns with a dual-path architecture on top of existing software-based graph *** first identify that memory accesses to the offset,edge,and state array have distinct locality and impact on *** then introduce the Skyway architecture,which consists of two primary components:1)a dedicated direct data path between the core and memory to transfer state array elements efficiently,and 2)a data-type aware fine-grained memory-side row buffer hardware for both the newly designed direct data path and the regular memory hierarchy data *** proposed Skyway architecture is able to improve the overall performance by reducing the memory access interference and improving data access efficiency with a minimal *** evaluate Skyway on a set of diverse algorithms using large real-world *** a simulated fourcore system,Skyway improves the performance by 23%on average over the best-performing graph-specialized hardware optimizations.

关键词： graph application computer architecture memory hierarchy

来源：评论

学校读者我要写书评

暂无评论

Morpho-contour exponential estimation algorithm for predicting breast tumor growth from MRI imagery

引用

International Journal of Information technology (Singapore) 2024年 1-16页

作者： Kaladevi, P. Kanimozhi, N. Nirmala, B. Sivasankari, R. Department of Computer Science and Engineering K. S. Rangasamy College of Technology (Autonomous) Tamil Nadu Tiruchengode 637215 India Department of Computational Intelligence School of Computing SRM Institute of Science and Technology Kattankulathur Tamil Nadu Chennai 603203 India Computer Science and Engineering Sri Ramachandra Faculty of Engineering and Technology Sri Ramachandra Institute of Higher Education and Research Porur Chennai 600116 India Department of Computer Science Engineering B. S. Abdur Rahman Crescent Institute of Science and Technology Vandalur Tamil Nadu Chennai 600 048 India

Breast tumor grows rapidly have an elevated risk of progressing to a metastatic state, leading to the spread of tumor tissues to other regions of the body. Early detection of such metastatic sites can be facilitated by predicting these tumors' growth rate. In this context, we introduce a novel Morpho-Contour Exponential Estimation (MoCEE) method to estimate the potential growth rate of identified breast tumors using magnetic resonance imaging (MRI) datasets. This methodology not only effectively captures morphological traces but also pinpoints regions indicative of infectious spread. This precision is achieved by enhancing the Mask R-CNN with an active contour function, ensuring the accurate segmentation of the breast tumor. Once segmented, the tumor's growth rate is computed using the exponential growth model, which accurately determines the rate by analyzing the distance between two distinct random points on the tumor's segmented region. Subsequently, this derived information, processed through a gradient boosting approach, predicts the likelihood of tumor growth across two different breast tumor categories. In our study, we assessed the proposed MoCEE method from two crucial angles. Firstly, we rigorously evaluated its ability to accurately segment both benign and malignant breast tumors. Our analysis involved comparing MoCEE's performance to that of various deep learning networks, yielding impressive segmentation accuracy percentages, such as 93.12% and 92.22% for benign and malignant tumors in BUSnet, and even higher values for other networks. Secondly, we explored MoCEE's capability to estimate tumor size, considering different probability growth rates (k^=0.2, 0.3, 0.4, and 0.5). Notably, proposed MoCEE proved highly accurate in predicting tumor growth rates across this range of probabilities. These findings highlight MoCEE's versatility and robustness, making it a valuable tool for precise breast tumor analysis and prediction. © The Author(s), under exclusiv

关键词： And gradient boosting approach Breast tumors Exponential growth model Mask R-CNN Morpho-contour exponential estimation (MoCEE)

来源：评论

学校读者我要写书评

暂无评论

Detection of Cyber-attacks in Food Industry using Multi-Layer Perceptron and Convolution Neural Network

Detection of Cyber-attacks in Food Industry using Multi-Laye...

引用

2023 International Conference on Sustainable Computing and Smart Systems, ICSCSS 2023

作者： Beulah Rani, I. Vinodh Ewards, S.E. Matthew Palmer, G. Kathrine, G.Jaspher W. Computer Science and Engineering Karunya Institute of Technology and Sciences Coimbatore India

ISBN: (纸本)9798350333602

Food insecurity refers to the lack of access to sufficient food, and this is a critical problem that many countries are currently tackling. While there are various causes contributing to this issue, cyberattacks are identified as the primary factor impacting the food industry after the concept of Industry 4.0 was introduced. The impact of cyberattacks on the food industry has been documented by several authors based on the technologies and the devices used in real-time, revealing the need for better detection methods. This study employs the use of three deep learning algorithms, multi-layer perceptron, Artificial neural network and convolution neural networks, to detect cyber-attacks based on the equipment used in food industry. The study achieves its objective by using multiple hidden layers trained for accurate detection. Although MLP has 96.04% accuracy CNN has the highest accuracy rate of 96.77%. © 2023 IEEE.

关键词： Multilayer neural networks

来源：评论

学校读者我要写书评

暂无评论

DSAFF-Net:A Backbone Network Based on Mask R-CNN for Small Object Detection

引用

computers, Materials & Continua 2023年第2期74卷 3405-3419页

作者： Jian Peng Yifang Zhao Dengyong Zhang Feng Li Arun Kumar Sangaiah Hunan Provincial Key Laboratory of Intelligent Processing of Big Data on Transportation Changsha University of Science and TechnologyChangsha410114China School of Computer and Communication Engineering Changsha University of Science and TechnologyChangsha410114China School of Computing Science and Engineering Vellore Institute of Technology(VIT)Vellore632014India

Recently,object detection based on convolutional neural networks(CNNs)has developed *** backbone networks for basic feature extraction are an important component of the whole detection ***,we present a new feature extraction strategy in this paper,which name is *** this strategy,we design:1)a sandwich attention feature fusion module(SAFF module).Its purpose is to enhance the semantic information of shallow features and resolution of deep features,which is beneficial to small object detection after feature fusion.2)to add a new stage called D-block to alleviate the disadvantages of decreasing spatial resolution when the pooling layer increases the receptive *** method proposed in the new stage replaces the original method of obtaining the P6 feature map and uses the result as the input of the regional proposal network(RPN).In the experimental phase,we use the new strategy to extract *** experiment takes the public dataset of Microsoft Common Objects in Context(MS COCO)object detection and the dataset of Corona Virus Disease 2019(COVID-19)image classification as the experimental object *** results show that the average recognition accuracy of COVID-19 in the classification dataset is improved to 98.163%,and small object detection in object detection tasks is improved by 4.0%.

关键词： Small object detection classification RPN MS COCO COVID-19

来源：评论

学校读者我要写书评

暂无评论

Depression Detection Using Deep Learning 11th

Depression Detection Using Deep Learning

引用

11th International Conference on Frontiers of Intelligent Computing: Theory and Applications, FICTA 2023

作者： Gopichand, G. Ramesh, Anirudh Tholappa, Vasant Sridara Pandian, G. School of Computer Science and Engineering Vellore Institute of Technology Vellore India

ISBN: (纸本)9789819967056

It is of utmost importance to take care of mental health. This is especially true in the ongoing COVID-19 pandemic. Depression is a crippling problem and must be treated quickly. It is useful to focus efforts on new, efficient, and accurate methods of depression detection. We intend to detect whether a person is in a depressed state of mind. We will be using deep learning models. The algorithms that have been used are feed-forward deep neural network (FNN), long short-term memory (LSTM) neural network, simple recurrent neural network (RNN), 1D convolutional neural network (1D CNN), and gated recurrent units (GRU) neural network. These models have been trained by utilizing electroencephalogram (EEG) sensor datasets. They classify individuals into either "depressed" or "non-depressed" categories. These models have been compared using confusion matrix, precision, recall score, F1 score, and accuracy. The best-performing models identified are FNN and 1D CNN. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

Classifying Skin Cancer and Acne using CNN 15

Classifying Skin Cancer and Acne using CNN

引用

15th International Conference on Knowledge and Smart technology, KST 2023

作者： Vasudeva, Kshitiza Chandran, Saravanan National Institute of Technology Department of Computer Science and Engineering Durgapur India

ISBN: (纸本)9781665477123

Several hospitals and dermatological clinics have adopted computer-vision-based diagnosis tools to aid in the early identification of skin cancer. The most frequent skin diseases are acne vulgaris and skin cancer. Acne Vulgaris affects 85% of the population in their lives, usually during adolescence. Benign Skin Cancer is the cancer commonly affecting people among all other types in developed and developing countries. To measure the success of medical treatment techniques, an objective evaluation of the lesion is required. Traditionally, dermatologists manually count the number of lesions by visual examination or scanning obtained photographs of the patient's skin and divide them into several categories. This old procedure is time intensive and necessitates a significant amount of work on the part of the physician. Using computer vision, automated the lesion detection, lesion classification, counting of Acne, counting of benign skin cancer and tracking of Acne Severity, making it simple for patients to analyse and track the results of their acne treatment. The goal of this study is to develop a Convolutional Neural network model to classify the lesions into acne and benign skin cancer. The proposed model is developed and trained on acne and different types of benign skin cancer images and achieved an accuracy of 96.4%. © 2023 IEEE.

关键词： computer vision

来源：评论

学校读者我要写书评

暂无评论

A Task Scheduling Approach for Optimizing 5G Edge Network Performance Through Cloud-Edge Collaborative Computing

A Task Scheduling Approach for Optimizing 5G Edge Network Pe...

引用

2023 International Conference on Evolutionary Algorithms and Soft Computing Techniques, EASCT 2023

作者： Shanthi, M.B. Rao, Vaishali K. P. Sudhakar, K.N. Parameshachari, B.D. Cmr Institute of Technology Department of Computer Science and Engineering Bengaluru India Cmr Institute of Technology Department of Information Science and Engineering Bengaluru India Rv University Department of Computer Science and Engineering Bengaluru India Nitte Meenakshi Institute of Technology Department of Electronics and Communication Engineering Bengaluru India

ISBN: (纸本)9798350313413

This document discusses various techniques that are crucial for 5G mobile communication. These techniques include high-frequency transmission, direct communication, device-to-device (D2D) communication and massive multiple input and multiple-output (MIMO) technologies. The document provides an overview of each technology, including its benefits and limitations. In addition to these techniques, the document also covers mobile edge computing (MEC), which involves distributing computer resources over mobile communication networks. The document outlines eight potential uses of MEC. Lastly, the document proposes a cloud-edge collaborative computing job scheduling system. The proposed technique aims to dynamically assign computer resources to edge devices based on network conditions and current workloads while considering network latency, computation capability, and energy use. Performance scores for the Cloud Only plan range from 4.0 to 4.5, which shows that the plan is very effective. Lower scores from 1.8 to 3.0 obtained by the Edge Only approach imply that its performance is constrained. © 2023 IEEE.

关键词： 5G technology Cloud Edge Resources Device-To-Device Communication Direct Communication High-Frequency Transmission Low Latency Massive MIMO Mobile Edge Computing Multiple-Carrier Techniques Task Scheduling

来源：评论

学校读者我要写书评

暂无评论

Classification of Traffic Signal Images Using Deep Neural Networks 1

引用

7th International Conference on Smart Trends in Computing and Communications, SmartCom 2023

作者： Kumar, S. Jitesh Santhanam, Naveen Jaisakthi, S.M. School of Computer Science and Engineering Vellore Institute of Technology Vellore India

ISBN: (数字)9789819907694

ISBN: (纸本)9789819907687

Traffic sign recognition system plays a crucial role in vehicles that assures the safety of humankind. The system provides feedbacks on-road information to the drivers on time which helps them in driving the car without accidents. Deep Neural Networks (DNN) are gaining more attention in the field of image analysis since it produces a high accuracy rate. Recently many researchers are focusing on the recognition of traffic signs using DNN to get better results. So, in this paper, we have applied the transfer learning technique to classify traffic signs. We choose to work with 2 architectures to classify road signs, namely, fine-tuned Convolutional Neural Network (CNN) and Resnet50. The two architectures were trained using the GTSRB—German Traffic Sign Recognition Benchmark downloaded from Kaggle. With these models, we have obtained good results of 98% and 75% respectively. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Roads and streets

来源：评论

学校读者我要写书评

暂无评论

Bus Dwell Time Forecasting using Machine Learning Models 7

Bus Dwell Time Forecasting using Machine Learning Models

引用

7th International Conference on Trends in Electronics and Informatics, ICOEI 2023

作者： Aswini, B.P. Savithramma, R.M. Sumathi, R. Siddaganga Institute of Technology Department of Computer Science and Engineering Tumakuru India

ISBN: (纸本)9798350397284

Bus Dwell Time (BDT) is a vital and most important contributing part of Bus Travel Time (BTT). Forecasting BDT is key for applications that predict the arrival time of buses at bus stops and intersections. With aim of aiding the BTT prediction, the objectives of the current work are to (i) Develop machine learning models using archived location data of buses to predict dwell times at all the bus stops of a route and (ii) Compare the performances of the models developed with a base model through error metrics. Multiple machine learning algorithms namely;k-Nearest Neighbors, Gradient Boosting and Extreme Gradient Boosting (XGBoost) regressor, Random Forest Regressor, and One- Dimensional Convolutional Neural Network models are identified and implemented based on the literature survey and compared with the Historic Average as the base model. The location logs of Tumakuru city service buses, Tumakuru, India are preprocessed, trained, and tested. Among the implemented models, XGBoost demonstrated better performance. The mean absolute percentage of errors demonstrated by the implemented models is above 20% indicating the need to improve the performance of the models as a future study. © 2023 IEEE.

关键词： Travel time

来源：评论

学校读者我要写书评

暂无评论

Balancing Techniques for Improving Automated Detection of Hate Speech and Offensive Language on Social Media 2

Balancing Techniques for Improving Automated Detection of Ha...

引用

2nd International Conference for Innovation in technology, INOCON 2023

作者： Reddy, B. Ajay Chandrasekhar Chandra, Girish Kumar Sisodia, Dilip Singh Anuragi, Arti National Institute of Technology Department of Computer Science and Engineering Raipur India

ISBN: (纸本)9798350320923

On social media networks like Twitter, Facebook, and Tumblr, people frequently share information. However, these platforms are also notorious for the spread of hate speech and insults, often posted anonymously. Hate speech involves using violent, abusive, or aggressive language towards a particular group based on factors such as gender, race, religion, or region. The prevalence of hate speech on these websites is a major concern, and manually detecting it can be time-consuming. To address this issue, this study presents an automated hate speech detection model that is evaluated on a publicly available Twitter dataset. The proposed method emphasizes data pre-processing, including stemming, term frequency-inverse document frequency (TF-IDF) for feature extraction, and various sampling techniques (random sampler, synthetic minority over-sampling technique (SMOTE), and ALL-KNN) to balance an imbalanced dataset. The logistic regression, support vector machine (SVM), and k-nearest neighbor (k-NN) machine learning classifiers were trained and tested using hold-out cross-validation to reduce overfitting and evaluate performance. The performance was evaluated using metrics such as accuracy, precision, and confusion matrix. The results showed that the logistic regression classifier using the SMOTE approach had the best performance, with an accuracy of 82%, a macro average of precision, recall, and an F1-score of 80%, 82%, and 79%, respectively. © 2023 IEEE.

关键词： Logistic regression

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：