检索结果-内蒙古大学图书馆

IEEE Symposium on High-Performance computer Architecture

作者： Jeongmin Hong Sungjun Cho Geonwoo Park Wonhyuk Yang Young-Ho Gong Gwangsun Kim Department of Computer Science and Engineering POSTECH Soongsil University School of Software

We propose overcoming the memory capacity limitation of GPUs with high-capacity Storage-Class Memory (SCM) and DRAM cache. By significantly increasing the memory capacity with SCM, the GPU can capture a larger fraction of the memory footprint than HBM for workloads that mandate memory oversubscription, resulting in substantial speedups. However, the DRAM cache needs to be carefully designed to address the latency and bandwidth limitations of the SCM while minimizing cost overhead and considering GPU's characteristics. Because the massive number of GPU threads can easily thrash the DRAM cache and degrade performance, we first propose an SCM-aware DRAM cache bypass policy for GPUs that considers the multi-dimensional characteristics of memory accesses by G PU s with SCM to bypass DRAM for data with low performance utility. In addition, to reduce DRAM cache probe traffic and increase effective DRAM BW with minimal cost overhead, we propose a Configurable Tag Cache (CTC) that repurposes part of the L2 cache to cache DRAM cacheline tags. The L2 capacity used for the CTC can be adjusted by users for adaptability. Furthermore, to minimize DRAM cache probe traffic from CTC misses, our Aggregated Metadata-In-Last-column (AMIL) DRAM cache organization co-locates all DRAM cacheline tags in a single column within a row. The AMIL also retains the full ECC protection, unlike prior DRAM cache implementation with Tag-And-Data (TAD) organization. Additionally, we propose SCM throttling to curtail power consumption and exploiting SCM's SLC/MLC modes to adapt to workload's memory footprint. While our techniques can be used for different DRAM and SCM devices, we focus on a Heterogeneous Memory Stack (HMS) organization that stacks SCM dies on top of DRAM dies for high performance. Compared to HBM, the HMS improves performance by up to 12.5× (2.9× overall) and reduces energy by up to 89.3% (48.1 % overall). Compared to prior works, we reduce DRAM cache probe and SCM write traffic by 91–9

关键词：

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Monte Carlo Denoising via Learning Contrastive Disentanglement Representation

Unsupervised Monte Carlo Denoising via Learning Contrastive ...

引用

IEEE International Conference on Multimedia and Expo (ICME)

作者： Xiao Liang Siyuan Duan Lijie Zheng Yuqian Zeng School of Computer Science and Software Engineering Southwest Petroleum University China

ISBN: (数字)9798350390155

ISBN: (纸本)9798350390162

Recent Monte Carlo denoising methods have achieved impressive progress in generating visually compelling images, which accelerate path-tracing based rendering. However, most of them rely on strong supervision, resulting in a high computational burden for creating paired data and potential overfitting issues. Instead, we present an unsupervised approach based on contrastive disentanglement representation. Specifically, we introduce disentanglement representation combined with cycle-consistency and adversarial loss to factorize noise and content features from a noisy input. Besides, we enforce contrastive learning on abundant samples generated in the backbone to regularize the extracted content distribution, reducing noise features. Moreover, we introduce a multi-scale modulation to refine the representation of auxiliary features, providing better denoising guidance. Experimental results demonstrate that our approach performs favor-ably against existing state-of-the-art unsupervised methods and generates comparable results against supervised models while requiring less inference time.

关键词： Monte Carlo methods Noise reduction Noise Modulation Contrastive learning Rendering (computer graphics) Feature extraction Visual effects Noise measurement

来源：评论

学校读者我要写书评

暂无评论

Gaussian Differential Privacy Integrated Machine Learning Model for Industrial Internet of Things

引用

SN computer science 2023年第5期4卷 454页

作者： Lazar, Arokia Jesu Prabhu Soundararaj, Sivaprakash Sonthi, Vijaya Krishna Palanisamy, Vishnu Raja Subramaniyan, Vanithamani Sengan, Sudhakar Department of Computer Science and Engineering AI & ML CMR Institute of Technology Telangana Hyderabad 501401 India Department of Software Systems School of Computer Science and Engineering Vellore Institute of Technology Tamil Nadu Vellore 632014 India Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Andhra Pradesh Vaddeswaram 522502 India Department of Computer Science and Engineering Kongu Engineering College Tamil Nadu Perundurai 638060 India Master of Computer Applications M. Kumarasamy College of Engineering Tamil Nadu Karur 638060 India Department of Computer Science and Engineering PSN College of Engineering and Technology Tamil Nadu Tirunelveli 627152 India

Agriculture, energy, mining, healthcare, and transportation are a few of the top industries transformed by the industrial internet of things (IIoT). Industry 4.0 mainly relies on machine learning (ML) to use the vast interconnectedness and large amounts of IIoT data that IIoT primarily drives. ML approaches are trained on confidential data generated by the IIoT environment often exposes privacy to adversarial assaults. Blockchain-based ML is established in the proposed secured IIoT research work to safeguard and enhance privacy. This research proposes a Gaussian differential privacy-integrated machine learning model (GDPIMLM), created for a scalable and controlled IIoT system, fully connected (FC) layer, implemented blockchain benefits to create a privacy-preserving mechanism while considering other limitations and reasonable time. A probability test using two altered Gaussian distributions (GD) forms the basis for defining Gaussian differential privacy (GDP). Due to a central restriction theorem for differential privacy, GDP is the primary privacy term within the family of Federal differential privacy (f-DP) policies. Ethereum is used to implement experimental evaluations. Data collections show that the recommended method enhances digital data privacy with industry-leading security without reducing performance. The performance analysis of the proposed model of secured industrial internet of things (IIoT) is studied using the accuracy of comparison with existing methods and for diverse distributed entities (DE). The proposed secured IIoT attains a minimal time consumption (TC) of 790 Sec—the highest accuracy of 92%, outperforming the existing method. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： Blockchain Industry Industry 4.0 Learning Privacy Security Smart objects

来源：评论

学校读者我要写书评

暂无评论

Fuzzy Deduplication Scheme Supporting Pre-verification of Label Consistency 1

引用

17th International Conference on Provable and Practical Security, ProvSec 2023

作者： Tang, Zehui Zeng, Shengke Li, Tao Cheng, Shuai Zheng, Haoyu School of Computer and Software Engineering Xihua University Chengdu610039 China State Key Laboratory of Public Big Data College of Computer Science and Technology Guizhou University Guiyang550025 China School of Computer Science Qufu Normal University Rizhao276825 China

ISBN: (数字)9783031455131

ISBN: (纸本)9783031455124

Efficiently and securely removing encrypted redundant data with cross-user in the cloud is challenging. Convergent Encryption (CE) is difficult to resist dictionary attacks for its deterministic tag. Server-aided mechanism is against such attacks while it may exist collusion. Focus on multimedia data, this paper proposes an efficient and secure fuzzy deduplication system without any additional servers. We also propose a notion of pre-verification of label consistency to compensate for the irreparable post-verification loss. Compared with other fuzzy deduplication schemes, our work has apparent advantages in deduplication efficiency and security based on a natural data set. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： Cryptography

来源：评论

学校读者我要写书评

暂无评论

The Early Detection of Dementia Disease Using Machine Learning Approach

The Early Detection of Dementia Disease Using Machine Learni...

引用

2023 International Conference on computer Communication and Informatics, ICCCI 2023

作者： Kabir, Md. Shariar Khanom, Jakia Bhuiyan, Md. Atik Tumpa, Zerin Nasrin Rabby, S.K. Fazlee Bilgaiyan, Saurabh Daffodil International University Dept. of Computer Science and Engineering Dhaka Bangladesh Daffodil International University Dept. of Software Engineering Dhaka Bangladesh Kalinga Institute of Industrial Technology School of Computer Engineering Bhubaneswar India

ISBN: (纸本)9798350348217

Early dementia detection is a crucial but challenging task in Bangladesh. Often, dementia is not recognized until it is too late to receive effective care. This results in part from a lack of knowledge about the illness and its signs and symptoms. Recent improvements in machine learning algorithms, however, may change this. In a recent study, we developed a model that can identify early dementia in Bangladesh using machine learning algorithms. This research paper proposed an efficient machine learning-based approach for early detection of dementia disease A dataset of 199 people with dementia and 175 healthy controls was used to develop the model. In 96% of the cases, the algorithm correctly identified dementia. This is a significant accomplishment that could revolutionize Bangladesh's dementia detection process. For patients to get the care they require, early dementia detection is essential. This study offers a proof-of-concept for the use of machine learning in dementia early detection & The results of this study suggest that machine learning models can be used as a powerful tool for early detection of dementia. © 2023 IEEE.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

CoSiNet: Dual-Branch Collaborative Siamese Network for Visual Object Tracking

CoSiNet: Dual-Branch Collaborative Siamese Network for Visua...

引用

International Conference on computer Supported Cooperative Work in Design

作者： Wenjun Zhou Yao Liu Nan Wang Yifan Wang Bo Peng School of Computer Science and Software Engineering Southwest Petroleum University China

ISBN: (数字)9798350349184

ISBN: (纸本)9798350349191

This article presents a dual-branch Collaborative Siamese network architecture designed for visual object tracking, which we refer to as CoSiNet. The dual-branch collaborative Siamese network comprises two network branches: a shallow branch, which focuses on target localization to enhance resistance to interference from objects with similar characteristics, and a deep branch, which emphasizes the extraction of more abstract semantic information related to the object. Furthermore, we have devised a Channel Attention Feature Enhancement Module and a Spatial Channel Attention Feature Enhancement Module to augment feature extraction while mitigating the influence of background noise. In the concluding stages, an Adaptive Fusion Module is employed to amalgamate the response maps from both branches, resulting in an enhanced final response map. Experimental results, conducted on two publicly available datasets, demonstrate that our algorithm outperforms other state-of-the-art techniques in terms of tracking performance.

关键词： Location awareness Resistance Visualization Semantics Collaboration computer architecture Interference

来源：评论

学校读者我要写书评

暂无评论

GAN-IVDS: An Intrusion Detection System for Intelligent Connected Vehicles Based on Generative Adversarial Networks 8

GAN-IVDS: An Intrusion Detection System for Intelligent Conn...

引用

8th International Conference on Data science in Cyberspace, DSC 2023

作者： Chen, Xiansen Xiao, Kun Luo, Lei Li, Yun Chen, Lirong School of Information and Software Engineering University of Electronic Science and Technology of China Chengdu China School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China Guangdong Weichen Information Technology Co. Ltd Dongguan China

ISBN: (纸本)9798350331035

This paper proposes an anomaly detection algorithm based on Generative Adversarial Networks (GANs) for intrusion detection in Intelligent Connected Vehicles (ICVs). To address the imbalance between normal network traffic and attack traffic in vehicular networks, as well as the difficulty in collecting attack data, we train the neural network solely with normal traffic. We employ the Gini index to perform data dimensionality reduction, selecting 34 features with the highest purity of information and enhancing model training speed via normalization functions. Our algorithm combines the generative capabilities of GANs and the reconstructive power of autoencoders, detecting abnormal network data by learning the distribution of normal network data, thus eliminating the need to analyze packet payloads and safeguarding user privacy. We also develop a vehicular network traffic data collection scheme to collect network traffic data from ICVs. Experimental results indicate that our algorithm demonstrates high accuracy in detecting anomalous network traffic in ICVs, with an accuracy rate of 94.9%. © 2023 IEEE.

关键词： Anomaly Detection Algorithm Generative Adversarial Networks Intelligent Connected Vehicles Intrusion Detection Network Security

来源：评论

学校读者我要写书评

暂无评论

Enhancing road safety with DL vision: Do driver distraction alerts hold the key?

引用

Neural Computing and Applications 2025年 1-21页

作者： Ali, Luqman Swavaf, Muhammad Alnajjar, Fady Zou, Zhao Mohan Ambali Parambil, Medha Mubin, Omar AlJassmi, Hamad Emirates Center for Mobility Research UAE University Abu Dhabi Al Ain United Arab Emirates Department of Civil and Environmental Engineering College of Engineering UAE University Abu Dhabi Al Ain United Arab Emirates Department of Computer Science and Software Engineering College of IT UAE University Abu Dhabi Al Ain United Arab Emirates School of Computer Data and Mathematical Sciences Western Sydney University SydneyNSW Australia

Driver distraction is a significant contributor to road accidents and traffic safety challenges. This study evaluates the performance of various YOLO models trained on a combination of custom and publicly available datasets, with their performance assessed using metrics such as mAP@50, precision, recall, F1-score, inference time, FPS, and model parameters. The aim is to determine the best trade-off between accuracy, speed, and computational efficiency, providing insights into the practical applicability of these models for distracted driving detection. Among the models, YOLOv8l achieves the highest accuracy with a mAP@50 of 0.921, precision of 0.911, recall of 0.895, and an F1-score of 0.899, making it suitable for scenarios prioritizing detection precision. YOLOv5m balances accuracy and efficiency, achieving a mAP@50 of 0.917 and an F1-score of 0.896, while maintaining moderate inference times. For real-time applications, YOLOv5n and YOLOv8n offer the fastest speeds, with inference times of 0.9 ms and FPS of 1111, at the cost of reduced accuracy. YOLOv8m provides a balance between performance and speed, achieving a mAP@50 of 0.915 and F1-score of 0.898, with an inference time of 3.4 ms and FPS of 294. A post-simulation survey reveals that participants find the visual warning system effective in mitigating driver distraction and enhancing perceived road safety. By integrating simulation data and user feedback, this study highlights the potential of advanced deep learning models like YOLO in improving driver assistance systems and promoting safer driving practices. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： Highway accidents

来源：评论

学校读者我要写书评

暂无评论

Bank Customer Churn Prediction based on Random Forest Algorithm 24

Bank Customer Churn Prediction based on Random Forest Algori...

引用

5th International Conference on computer Information and Big Data Applications, CIBDA 2024

作者： Zhang, Nini Zheng, Yuzhi Duan, Chao Software Engineering Institute of Guangzhou Guangzhou510000 China School of Computer Science South China Normal University Guangzhou510000 China Guangzhou Xinhua University Guangzhou510000 China

ISBN: (纸本)9798400718106

Customer churn is a key concern for service industries such as banking, telecommunications and insurance. If they can know in advance which users have a higher probability of churn, institutions can formulate corresponding plans to restore them. This study is based on the random forest model and builds a credit card customer churn classification model by analyzing credit card customers’ historical transactions, credit card usage and other related characteristics. Use the data set on the kaggle platform for training and verification, and evaluate the performance of the model through cross-validation. This model can predict whether there is a risk of customer churn based on customer characteristics. Through the verification and optimization of the model, the accuracy and robustness of the model are improved. Experimental results show that the algorithm based on the random forest model outperforms the decision tree and support vector machine algorithms in predicting credit card customer churn, and its time complexity outperforms the neural network algorithm. © 2024 ACM.

关键词： Decision trees

来源：评论

学校读者我要写书评

暂无评论

A Ginkgo Detection Algorithm in Complex Environments Based on Improved YOLOv7

A Ginkgo Detection Algorithm in Complex Environments Based o...

引用

6th International Conference on Intelligent Computing and Signal Processing (ICSP)

作者： Chi Ma Qiang Guo School of Computer Science and Engineering Huizhou University Huizhou China School of Computer Science and Software Engineering University of Science and Technology LiaoNing Anshan China

ISBN: (数字)9798350376548

ISBN: (纸本)9798350376555

Detections of Ginkgoes are prerequisites for later counting and harvesting. Due to the uneven distribution of samples, the detection speed and accuracy of existing algorithms cannot adapt to the impact of complex environments. Therefore, an enhanced model YOLOv7-DC based on YOLOv7 is proposed in this paper, which redesigns the detection network and introduces a new feature fusion method. DCNv2 is embedded in the efficient layer aggregation network (ELAN), while PConv is utilized instead of conventional convolution to reduce the parameter impact of DCNv2. Moreover, the attention mechanism CBAM is introduced during training to enhance spatial and channel information, and the ConvMixer architecture is employed to capture spatial and channel relationships within the features, which are transmitted to the detection head through attention mechanism, improving the model's detection accuracy for each specific classification sample. Experimental results show that our YOLOv7-DC achieves both excellent detection speed and recognition rate in various classification tasks. The improved model's average detection accuracy is increased by 6.2% compared to previous algorithms, and the model parameters are reduced by 13%. It is proved that YOLOv7-DC is more suitable for scenarios with imbalanced samples and complex environments.

关键词： Training Accuracy Attention mechanisms Convolution Signal processing algorithms Lighting Medical services Feature extraction Environmental factors Detection algorithms

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：