检索结果-内蒙古大学图书馆

A Hybrid Approach to Dynamic Thread Allocation Driven by Workload and Task Prioritization

SN computer Science 2025年第3期6卷 1-19页

作者： Abirami, K. Harini, N. Department of Computer Science and Engineering Amrita School of Computing Coimbatore Amrita Vishwa Vidyapeetham Coimbatore India

In today’s rapidly changing world, cloud service providers face numerous challenges in managing resources and meeting customer demands. To address these challenges, cloud service providers should prioritize the tasks and allocate threads efficiently. By allocating threads efficiently, service providers can optimize resource utilization and reduce response times and job completion times which in turn improves customer satisfaction. We propose a hybrid model that integrates the threshold-driven thread Balancer (TDTB) for dynamic thread allocation with a Customer Retention Model (CRM) to prioritize tasks based on customer behaviour. Our model initiates the establishment of a thread pool configured with a dynamic threshold based on workload (on-demand requests). Concurrently, a queue to handle incoming customer requests is set up alongside a CRM priority queue designed for task prioritization. Our model calculates Job duration and completion times for n priority classes. Crown consumers experience shorter durations and completion times compared to other priority classes. Tasks are completed in descending order of priority. The results of thorough experimentation clearly show that crown consumer tasks are highly satisfied. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Customer segmentation Response time Service differentiation Thread allocation Throughput Workload

来源：评论

学校读者我要写书评

暂无评论

Generating Markov Logic Networks Rulebase Based on Probabilistic Latent Semantics Analysis

引用

Tsinghua Science and Technology 2023年第5期28卷 952-964页

作者： Shan Cui Tao Zhu Xiao Zhang Liming Chen Lingfeng Mao Huansheng Ning School of Computer and Communication Engineering University of Science and Technology BeijingBeijing 100083China Computer School University of South ChinaHengyang 421001China School of Computing Ulster UniversityBelfastBT151APUK

Human Activity Recognition(HAR)has become a subject of concern and plays an important role in daily *** uses sensor devices to collect user behavior data,obtain human activity information and identify *** Logic Networks(MLN)are widely used in HAR as an effective combination of knowledge and *** can solve the problems of complexity and uncertainty,and has good knowledge expression ***,MLN structure learning is relatively weak and requires a lot of computing and storage ***,the MLN structure is derived from sensor data in the current *** that the sensor data can be effectively sliced and the sliced data can be converted into semantic rules,MLN structure can be *** this end,we propose a rulebase building scheme based on probabilistic latent semantic analysis to provide a semantic rulebase for MLN *** a rulebase can reduce the time required for MLN structure *** apply the rulebase building scheme to single-person indoor activity recognition and prove that the scheme can effectively reduce the MLN learning *** addition,we evaluate the parameters of the rulebase building scheme to check its stability.

关键词： Markov Logic Network(MLN) structure learning rulebase construction probabilistic latent semantics

来源：评论

学校读者我要写书评

暂无评论

DSAF: A Dual-Stage Attention Based Multimodal Fusion Framework for Medical Visual Question Answering

引用

SN computer Science 2025年第4期6卷 1-16页

作者： Mukesh, K. Jayaprakash, S.L. Prasanna Kumar, R. Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Tamilnadu Chennai 601103 India

In the realm of clinical healthcare, medical visual question answering systems emerge as a pivotal innovation that plays a crucial role in clinical decision-making and patient care. They are designed to interpret medical images in response to free-form natural language questions, thereby facilitating a more nuanced and comprehensive understanding of patient diagnostics. This approach seeks to bridge the gap between complex medical imagery and the clinical questions they prompt, offering a novel avenue for enhancing diagnostic accuracy and efficiency. We introduce the Dual Stage Attention-based Fusion framework, which leverages an advanced feature fusion strategy, along with image encoding techniques and state-of-the-art biomedical language models, to process and analyze medical images and textual queries concurrently. By employing a dual-stage attention mechanism, the proposed model enhances feature extraction and integration, enabling a more accurate and refined analysis of both visual and textual data. The model specifically addresses the challenges commonly associated with medical visual question answering, including the intricate nature of medical imagery and the scarcity of high-quality, annotated medical datasets. The proposed multimodal augmentation technique, involving an autoencoder and generative language model, significantly enriches the data for model training which mitigates the data scarcity issue and enhances the model’s ability to generalize across different medical queries. Empirical validation of our model on the VQA-RAD and SLAKE datasets demonstrates its exceptional performance, achieving state-of-the-art accuracies of 80.8% and 86.2% respectively. These results underscore the model’s potential in AI-enhanced medical diagnostics and decision support systems to offer unprecedented diagnostic insights and patient care solutions. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Data fusion Dual-stage attention Multimodal learning Vision language model Visual reasoning

来源：评论

学校读者我要写书评

暂无评论

Adaptive algorithm for controlling the power management system on offshore jack-up drilling rigs

引用

Energy Sources, Part A: Recovery, Utilization and Environmental Effects 2024年第1期46卷 16717-16737页

作者： Cemeljic, Hrvoje Havelka, Juraj Jeremic, Aleksandar Kuzle, Igor Department of Energy and Power Systems Faculty of Electrical Engineering and Computing University of Zagreb Zagreb Croatia Department of Electrical & Computer Engineering McMaster School of Biomedical Engineering McMaster University HamiltonON Canada Faculty of Electrical Engineering and Computing University of Zagreb Zagreb Croatia

The power systems of offshore jack-up drilling rigs consist of diesel generators running in parallel load-sharing mode, controlled by an automatic Power Management System (PMS). In this paper, the operational performance of the diesel generators (DG) and the PMS on a jack-up drilling rig is investigated, focusing on two critical offshore drilling operations: "pipe tripping" and "pulling/pumping out of the hole" (POOH). During these operations, large power swings occur in the system, subjecting the DGs to sudden load surges. These severe imbalances in the system cause the PMS to intermittently start and stop the engines throughout the operation, leading to various problematic conditions in the power plant. The paper provides a novel solution for intermittent starting and stopping of the DGs in the form of development and implementation of an adaptive PMS algorithm, a mathematical model of the system and a machine learning approach to event prediction of PMS operation in offshore drilling. Research demonstrates a noteworthy decrease in intermittent DG starting and stopping when the adaptive algorithm is implemented, with a 96% reduction during pipe tripping and an 87% reduction during POOH. This improvement comes at the expense of an 11% increase in engine running hours and a 1.5% rise in fuel consumption. Furthermore, by configuring the adaptive algorithm to economy mode, savings of up to 1700 liters of fuel per month are achievable. © 2024 Taylor & Francis Group, LLC.

关键词： Adaptive algorithms

来源：评论

学校读者我要写书评

暂无评论

SAM struggles in concealed scenes—empirical study on “Segment Anything”

引用

Science China(Information Sciences) 2023年第12期66卷 278-280页

作者： Ge-Peng JI Deng-Ping FAN Peng XU Bowen ZHOU Ming-Ming CHENG Luc VAN GOOL School of Computing Australian National University Computer Vision Lab(CVL) ETH Zurich Department of Electronic Engineering Tsinghua University College of Computer Science Nankai University

Large models open up new opportunities for artificial intelligence. In the past few months, there has been a boom in training foundation models on the vast linguistic corpus to produce amazing applications, e.g., Chat GPT, *** natural language processing and multimodal learning communities have been revolutionized. Large models' capacity for generalization and emergent makes it easy for users to believe that large models can solve anything.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Enhancing Secure Data Transmission Through Deep Learning-Based Image Steganography and Super-Resolution Generative Adversarial Networks

引用

SN computer Science 2024年第8期5卷 1-15页

作者： Ramesh, Sreyas Sukanth, B.N. Jaswanth, Sathyavarapu Sri Singh, Rimjhim Padam Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India

In the era of multimedia technology digital images are essential and keeping them safe from unauthorised access is crucial. To address this issue, the proposed research explores the intersection of image steganography and deep learning for enhanced secure data transmission. It proposes a unique approach that leverages skip-connection-based residual network to extract semantic features and reconstruct images for secure data transmission. This paper introduces a novel steganographic approach combining LSB embedding with sprite image generation instead of directly using image features. Hence, the proposed method effectively hides data within images while maintaining high invisibility. Furthermore, Super-Resolution Generative Adversarial Network (SRGAN) is employed to improve the quality of the reconstructed images. Several other custom and state-of-art convolutional architectures have also been investigated for semantic feature extraction and image reconstruction, including Convolutional Autoencoders (CAE), Deep convolutional network (DCN), DenseNet, InceptionNet and the XceptionNet model, etc. for the comprehensive comparisons against the proposed model. The proposed technique proved to be the best performing model with an SSIM value of 1.00 and PSNR value of 33.04 on the test data. Overall, this research work combines deep learning with steganography to create a secure and hidden communication method. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Data security Least significant bit Sprite image Steganography Super-resolution generative adversarial network

来源：评论

学校读者我要写书评

暂无评论

Convergence of Deep Learning and Forensic Methodologies Using Self-attention Integrated EfficientNet Model for Deep Fake Detection

引用

SN computer Science 2024年第8期5卷 1-12页

作者： Singh, Rimjhim Padam Sree, Nichenametla Hima Reddy, Koti Leela Sai Praneeth Jashwanth, Kandukuri Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India

The immensely increasing number of Deepfake technologies poses significant challenges to digital media integrity, leading to the immediate need for effective Deepfake detection methods. In light of the growing threat posed by sophisticated synthetic media, this paper tackles the urgent need for accurate deepfake detection in real-time by proposing a novel classification model employing a fine-tuned EfficientNet-B0 model with an optimally integrated Self-attention mechanism. The study also analyzes various Deepfake generation techniques by utilizing a suite of Convolutional Neural Networks namely, XceptionNet, NasNetLarge, InceptionV3, ConNextv1, ConvNext2 and EfficientB0 model on the Celeb-DF dataset for baseline model selection and performance comparison purposes. The results prove that the proposed fine-tuned EfficientNet-B0 model with integrated self-attention mechanism is the most superior at identifying complex long-range dependencies and relations in various unconnected image regions, highlighting the importance of attention mechanisms in improving deep learning models for multimedia forensics with a remarkable F-score of 97.67%. In an era marked by the proliferation of sophisticated artificial intelligence (AI)-driven synthetic media, these findings establish the foundation for strengthening digital media platform security and protecting information authenticity. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Celeb-DF dataset EfficientNetB0 Multimedia forensics NasNetLarge Self-attention mechanism Xception Deepfake detection

来源：评论

学校读者我要写书评

暂无评论

A Multi-path Droplet Routing Protocol for Digital Microfluidic Biochip

引用

Informatica (Slovenia) 2024年第7期48卷 95-102页

作者： Swain, Jyotiranjan Pyne, Sumanta School of Computing and Information Technology REVA University Bangalore India Department of Computer Science and Engineering National Institute of Technology Rourkela India

Digital microfluidic biochip provides an alternative platform to synthesize the biochemical protocols. Droplet routing in biochemical synthesis involves moving multiple droplets across the biochip simultaneously. It involves challenges like moving droplets without violating the fluid constraints. This article proposed a multi-path droplet routing protocol. Multiple routes are explored and validated using special packets. These routes are then classified based on user-defined heuristics. A single route is sequentially selected for each droplet. The routes are then compacted to generate a parallel moving sequence. The simulation result shows 2.087% and 4.952% improvement in the latest arrival time for free and virtual topology, respectively. © 2024 Slovene Society Informatika. All rights reserved.

关键词： Biochips

来源：评论

学校读者我要写书评

暂无评论

Deep ensemble model for skin cancer classification with improved feature set

引用

Multimedia Tools and Applications 2025年第10期84卷 7599-7626页

作者： Burada, Sreedhar Manjunathswamy, B.E. Kumar, M. Sunil Department of Computer Science and Engineering Don Bosco Institute of Technology Karnataka Bengaluru560074 India Department of Computer Science and Engineering School of Computing Mohan Babu University Sree Vidyanikethan Engineering College Andhra Pradesh Tirupati517102 India

Early identification of skin cancer is mandatory to minimize the worldwide death rate as this disease is covering more than 30% of mortality rates in young and adults. Researchers are in the move of proposing advanced models for the accurate classification of skin disease. This paper introduces a novel skin cancer detection model under the stages of preprocessing, feature extraction, classification as well as segmentation. The given input is first preprocessed via Gaussian filtering. Subsequently, the Deep learning model termed as ‘U-net’ is used for the segmentation process. Afterward, the extraction of the feature set including features like Improved Center symmetric Local Binary Pattern (ICS-LBP), Local Gradient Pattern (LGP), Median binary pattern (MBP) is performed. Classification takes place using Deep Ensemble Model based on the extracted features, that includes BiGRU, Deep Maxout, and Improved Deep Convolutional Neural Network (CNN), respectively. Also, as the hyperparameter tuning of the classification model enhances the performance, this work introduces a new Hybrid Atom and Arithmetic Operation Algorithm (HAAOA) algorithm for tuning the optimal weights of the Bi-GRU (Bi-directional Gated Recurrent Unit) and Deep Maxout model. At last, the validation of the proposed work is performed through different performance measures by comparing it with different algorithms. Accordingly, accuracy of the adopted method at 90th learning percentage is 93.56%, which is (~) 82.63%, 85.15%, 84.98%, 83.66% and 83.42% better than AOA, ASO, BWO, BMO and DO, respectively. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Monocular 3D object detection with Pseudo-LiDAR confidence sampling and hierarchical geometric feature extraction in 6G network

引用

Digital Communications and Networks 2023年第4期9卷 827-835页

作者： Jianlong Zhang Guangzu Fang Bin Wang Xiaobo Zhou Qingqi Pei Chen Chen School of Electronic Engineering Xidian UniversityXi'an710071China School of Computer Science and Technology College of Intelligence and ComputingTianjin UniversityTianjin300000China School of Telecommunications Engineering Xidian UniversityXi'an710071China

The high bandwidth and low latency of 6G network technology enable the successful application of monocular 3D object detection on vehicle *** 3D-object-detection-based Pseudo-LiDAR is a low-cost,lowpower solution compared to LiDAR solutions in the field of autonomous ***,this technique has some problems,i.e.,(1)the poor quality of generated Pseudo-LiDAR point clouds resulting from the nonlinear error distribution of monocular depth estimation and(2)the weak representation capability of point cloud features due to the neglected global geometric structure features of point clouds existing in LiDAR-based 3D detection ***,we proposed a Pseudo-LiDAR confidence sampling strategy and a hierarchical geometric feature extraction module for monocular 3D object *** first designed a point cloud confidence sampling strategy based on a 3D Gaussian distribution to assign small confidence to the points with great error in depth estimation and filter them out according to the ***,we present a hierarchical geometric feature extraction module by aggregating the local neighborhood features and a dual transformer to capture the global geometric features in the point ***,our detection framework is based on Point-Voxel-RCNN(PV-RCNN)with high-quality Pseudo-LiDAR and enriched geometric features as *** the experimental results,our method achieves satisfactory results in monocular 3D object detection.

关键词： Monocular 3D object detection Pseudo-LiDAR Confidence sampling Hierarchical geometric feature extraction

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：