检索结果-内蒙古大学图书馆

2024 International Conference on Image Processing and Artificial Intelligence, ICIPAl 2024

作者： Jia, Huiying Su, Xuesong School of Computer Science and Engineering Tianjin University of Technology Tianjin China Technology Inspection Center of Shengli Oilfield SINOPEC Shandong Dongying China

ISBN: (纸本)9781510681514

This study proposes an end-to-end method aimed at achieving intelligent colorization of images. This approach enables automatic colorization of grayscale images by taking in a grayscale image and a colored reference image as input. We designed a deep neural network model to learn the mapping relationship between grayscale images and colored reference images in an end-to-end manner. During training, we utilized a large amount of image data with real color labels and optimized the model using generative adversarial networks. Experimental results demonstrate the superiority of our method in color restoration and detail preservation across different types of images. Compared to traditional methods, the end-to-end approach offers higher levels of automation and faster colorization speed, opening up new possibilities in the field of image colorization. © 2024 SPIE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Internet of Intelligent Things (IoT) - An Overview 1

Internet of Intelligent Things (IoT) - An Overview

引用

1st International Conference on Computational Intelligence, Communication Technology and Networking, CICTN 2023

作者： Salama, Ramiz Al-Turjman, Fadi Aeri, Manisha Yadav, Satya Prakash Ai and Robotics Institute Research Center for Ai and IoT Near East University Nicosia Department of Computer Engineering Mersin10 Turkey Ai and Robotics Institute Near East University Artificial Intelligence Engineering Dept. Mersin Nicosia10 Turkey University of Kyrenia Research Center for Ai and IoT Faculty of Engineering Kyrenia Mersin10 Turkey Graphic Era Hill University Computer Science and Engineering Dehradun India Department of Computer Science and Engineering Greater Noida201306 India

ISBN: (纸本)9798350338027

The basic goal of the IoT is to permit the online connection and communication of everyday physical objects like appliances, vehicles, and other gadgets. This paves the way for effortless and automated data collection and exchange, as well as interaction between the objects themselves and with humans. The Internet of Everything (IoE) is an expansion of the Internet of Things (IoT) that aims to maximize and improve human life by connecting and making the most of all available resources (humans, information, and technology). The transportation, healthcare, agricultural, energy, and manufacturing sectors are all examples. The IoIT aims to incorporate AI and ML within the IoT's interconnected network of physical items. These objects can then independently behave in response to the information they have gathered and the instructions they have been given. This has the potential to lead to smarter, more efficient systems that can react to new information and human input in real time. The Internet of Things, the Internet of Everything, and the Industrial Internet of Things all have the same overarching goal: to pave the way for the development of more efficient and intelligent systems and processes that can enhance a wide range of human endeavors. However, in order to make sure that these technologies are used in a responsible and ethical way, it will be necessary to carefully analyze their consequences and the issues they provide. © 2023 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Practical Deep Learning Models for QIM-based VoIP Steganalysis 24

Practical Deep Learning Models for QIM-based VoIP Steganalys...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Zhang, Cheng School of Computer Science and Technology Engineering Research Center of Mine Digitalization China University of Mining and Technology Ministry of Education Jiangsu Xuzhou China

ISBN: (纸本)9798400706868

Quantization index modulation (QIM) based VoIP steganography can conceal secret information in VoIP streams. Malicious users could use this technology to conduct illegal activities, threatening network and public security. Hence, practical steganalysis models that could detect QIM-based VoIP steganography are urged to be developed. In recent years, deep learning (DL) models have been investigated for this task, and exciting outcomes have been achieved. However, existing models are far from practical. Two major challenges are required to be addressed. First, there is still significant room for improvement in detection accuracy. Second, studies that balance the detection accuracy and response time are still insufficient. In this context, our main research topic fits in the QIM-based VoIP steganalysis theme, which aims to detect QIM-based steganography in VoIP streams in a fast and accurate manner. © 2024 ACM.

关键词： covert channel

来源：评论

学校读者我要写书评

暂无评论

CEMA-LSTM:Enhancing Contextual Feature Correlation for Radar Extrapolation Using Fine-Grained Echo Datasets

引用

computer Modeling in engineering & sciences 2023年第4期135卷 45-64页

作者： Zhiyun Yang Qi Liu HaoWu Xiaodong Liu Yonghong Zhang School of Computer and Software Engineering Research Center of Digital ForensicsMinistry of EducationNanjing University of Information Science and TechnologyNanjing210044China School of Computing Edinburgh Napier UniversityEdinburghEH105DTUK School of Automation Nanjing University of Information Science TechnologyNanjing210044China

Accurate precipitation nowcasting can provide great convenience to the public so they can conduct corresponding arrangements in advance to deal with the possible impact of upcoming heavy *** relevant research activities have shown their concerns on various deep learning models for radar echo extrapolation,where radar echo maps were used to predict their consequent moment,so as to recognize potential severe convective weather ***,these approaches suffer from an inaccurate prediction of echo dynamics and unreliable depiction of echo aggregation or dissipation,due to the size limitation of convolution filter,lack of global feature,and less attention to features from previous *** address the problems,this paper proposes a CEMA-LSTM recurrent unit,which is embedded with a Contextual Feature Correlation Enhancement Block(CEB)and a Multi-Attention Mechanism Block(MAB).The CEB enhances contextual feature correlation and supports its model to memorize significant features for near-future prediction;the MAB uses a position and channel attention mechanism to capture global features of radar *** practical radar echo datasets were used involving the FREM and CIKM 2017 *** quantification and visualization of comparative experimental results have demonstrated outperformance of the proposed CEMA-LSTMover recentmodels,e.g.,PhyDNet,MIM and PredRNN++,*** particular,compared with the second-rankedmodel,its average POD,FAR and CSI have been improved by 3.87%,1.65%and 1.79%,respectively on the FREM,and by 1.42%,5.60%and 3.16%,respectively on the CIKM 2017.

关键词： Radar echo extrapolation attention mechanism long short-term memory deep learning

来源：评论

学校读者我要写书评

暂无评论

Machine Learning for Climate Change Impact Assessment and Adaptation Planning 2

Machine Learning for Climate Change Impact Assessment and Ad...

引用

2nd IEEE International Conference on Trends in Quantum Computing and Emerging Business Technologies, TQCEBT 2024

作者： Balakrishnan, T. Suresh Krishnan, Prabhakar Ebenezar, U. Samson Mohammed Nizarudeen, M. Kamal, N. Saveetha Institute of Medical and Technical Sciences Saveetha University Saveetha School of Engineering Department of Computer Science and Engineering Chennai India Amrita Vishwa Vidyapeetham Center for Cyber Security Systems and Networks Amritapuri-Campus India Saveetha Institute of Medical and Technical Sciences Saveetha University Saveetha School of Engineering Department of Computer Science and Engineering Tamil Nadu Chennai India Care College of Engineering Department of Computer Science and Engineering Tamil Nadu Trichy India Grt Institute of Engineering and Technology Dept. of Computer Science and Engineering Tamil Nadu Tiruttani India

ISBN: (纸本)9798350384277

The presented research introduces a new advanced data analytics methodology for climate change impact assessment and adaptive planning that utilizes machine learning and deep-learning techniques. Quantitative analysis was performed over several databases such as GHCN, ERA5, National centers for Environmental Prediction (NCEP) and JRA-55 in addition to CRU for a thorough assessment of the proposed methodology. In climate modeling, the Convolutional Neural Networks (CNN) showed better accuracy than ERA5 and JRA-55 which outperformed Global Historical Climatology Network (GHCN) with lower Mean Absolute Error (MAE), Root Mean Squared Error values. The Early Warning System, which was built using the LSTM network algorithm, had a higher accuracy of 0.90 and recall rate of 0.88 especially with JRA-55 proving its effectiveness in early detection for extreme weather events before they occur. The RL-based adaptive planning component performed well across all datasets, with ERA5 reaching the best feedback loop performance score of 94%. The precision agriculture optimization, which was enabled by Random Forest revealed different accuracy and AUC-ROC values across various datasets showing flexibility. Against international climate agreements, the overall alignment score was 0.78 with ERA5 being slightly better at 0.80 using Support Vector Machines (SVM). These quantitative assessments accentuate the practicality, efficacy, and applicability of this methodology in various datasets making it an apt tool for evidence-based decisions on climate resilience strategies at the global level. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Implementing Machine Learning Algorithms to Indentify Distributed Denial-of-Service Attacks 7

Implementing Machine Learning Algorithms to Indentify Distri...

引用

7th International Conference on Image Information Processing, ICIIP 2023

作者： Mehra, Ankush Singh, Gurpreet Badotra, Sumit Kaur, Balwinder Verma, Amit Lovely Professional University Department of Computer Science And Engineering Punjab Phagwara India Bennett University School of Computer Science Engineering And Technology Greater Noida India University Center For Research And Development Chandigarh University Department of Computer Science And Engineering Mohali Gharuan India

ISBN: (纸本)9798350371406

An Internet server, service, or network may be the subject of a distributed denial-of-service (DDoS) assault, in which the attacker attempts to interrupt regular traffic by flooding the target with an excessive amount of data requests. Network security has been severely compromised as a result. We choose the vast majority of network parameters to characterize the traffic in order to get a complete picture of it. As a further step toward lowering the detection time complexity, we employ the PCA approach to compress the features' dimensions. Using principal component analysis (PCA), we can save much of the original data while drastically cutting down on forecast time. After data dimensionality reduction, the RNN is used to train and get a detection model. In comparison to other approaches for detecting DDoS assaults, PCA-RNN performs much better on the real-world dataset, as shown by the evaluation results. © 2023 IEEE.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

A Novel Gas Classification Method Based on The Improved GasNet 43

A Novel Gas Classification Method Based on The Improved GasN...

引用

43rd Chinese Control Conference, CCC 2024

作者： Qin, Xiaolin Cheng, Lei Tao, Xiang Engineering Research Center for Metallurgical Automation and Measurement Technology of Ministry of Education Wuhan430081 China Wuhan University of Science and Technology Institute of Robotics and Intelligent Systems Wuhan430081 China Wuhan University of Science and Technology School of Information Science and Engineering Wuhan430081 China

ISBN: (纸本)9789887581581

Accurately and promptly identifying harmful gases in the environment can effectively prevent many accidents. Most existing pattern recognition methods are focused on addressing gas classification problems in static scenarios, but the real world is a dynamic and changing environment. Therefore, researching gas classification in continuous learning scenarios will better meet the needs of suddenly emerging gas classification tasks in real-world environments. To address this issue, an incremental learning model was introduced into gas classification, and an improved deep convolutional network was proposed for gas image feature extraction. The proposed model and other baseline methods experimented on an open-source gas dataset. The proposed model reduces knowledge loss due to catastrophic forgetting and effectively classifies the 10 gas samples in the dataset in continuous learning scenarios compared to non-incremental methods. It outperforms other baseline incremental methods by optimizing the network's shortcut connections, making the feature extraction network easier to train and improving gas classification performance. In dynamic learning scenarios, the proposed model achieved a classification accuracy of 86.30% for 10 gas samples. The experimental results demonstrate that the proposed method effectively mitigates catastrophic forgetting interference and completes gas classification tasks in continuous learning scenarios. © 2024 Technical Committee on Control Theory, Chinese Association of Automation.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Deep Reinforcement Learning for AoI-aware Transmission Control in Energy-Harvesting Grant-Free NOMA IoT Networks 11

Deep Reinforcement Learning for AoI-aware Transmission Contr...

引用

11th International Symposium on Telecommunication, IST 2024

作者： Sheikhi, Marzieh Hakami, Vesal Zarrabi, Houman Iran University of Science and Technology School of Computer Engineering Tehran Iran Iran Telecommunication Research Center Department of Communication Technologies Tehran Iran

ISBN: (纸本)9798350356250

The advent of Internet of Things (IoT) networks has revolutionized various real-time applications, such as healthcare, smart cities, and remote automation, by enabling seamless data exchange and real-time monitoring and control. Real-time applications require fresh data, which is quantified by a novel metric, Age of Information (AoI), defined as the time elapsed since the generation time of last successfully received data. Additionally, due to energy limitations in most IoT scenarios, the trade-off between age and energy becomes an important challenge in these networks. Motivated by the need to optimize both data freshness and energy efficiency, this paper considers a real-time scenario in energy-harvesting (EH) IoT networks that operate based on grant-free (GF) Non-Orthogonal Multiple Access (NOMA) mechanism. Given the foresighted nature of the transmission control problem due to the stochastic dynamics of the AoI and energy charging processes, we introduce a deep reinforcement learning (DRL) framework that dynamically manages transmission decisions and resource block allocation, minimizing the average sum AoI across IoT devices, while considering energy feasibility. Through extensive simulations, we demonstrate that our proposed DRL-based approach significantly outperforms existing methods in terms of expected cumulative cost, showcasing its potential for enhancing the performance of future IoT networks. © 2024 IEEE.

关键词： Resource allocation

来源：评论

学校读者我要写书评

暂无评论

An image inpainting method based on the high-efficiency transformer

An image inpainting method based on the high-efficiency tran...

引用

2024 International Conference on Image Processing and Artificial Intelligence, ICIPAl 2024

作者： Yang, Jiaqi Su, Xuesong School of Computer Science and Engineering Tianjin University of Technology Tianjin China Technology Inspection Center of Shengli Oilfield SINOPEC Shandong Dongying China

ISBN: (纸本)9781510681514

In this paper, we introduce an image restoration network that is based on the High-Efficiency Transformer (HET). The model utilizes fast Fourier convolution for image enhancement, incorporates self-attention mechanisms for capturing contextual information, and employs convolution for feature extraction and classification. Additionally, we improve the network's performance and convergence speed by implementing batch normalization and residual concatenation. Experimental results show substantial enhancements in image restoration achieved through our deep learning-based approach. The reconstructed images demonstrate improved clarity, enhanced details, and reduced dependence on manual interventions. © 2024 SPIE.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Control Flow Divergence Optimization by Exploiting Tensor Cores 24

Control Flow Divergence Optimization by Exploiting Tensor Co...

引用

61st ACM/IEEE Design Automation Conference, DAC 2024

作者： Pang, Weiguang Jiang, Xu Liu, Songran Qiao, Lei Fu, Kexue Gao, Longxiang Yi, Wang University of Electronic Science and Technology of China China Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China School of Computer Science and Engineering Northeastern University China Beijing Institute of Control Engineering Beijing China Uppsala University Sweden

ISBN: (纸本)9798400706011

Kernels are scheduled on Graphics Processing Units (GPUs) in the granularity of GPU warp, which is a bunch of threads that must be scheduled together. When executing kernels with conditional branches, the threads within a warp may execute different branches sequentially, resulting in a considerable utilization loss and unpredictable execution time. This problem is known as the control flow divergence. In this work, we propose a novel method to predict threads' execution path before the launch of the kernel by deploying a branch prediction network on the GPU's tensor cores, which can efficiently parallel run with the kernels on CUDA cores, so that the divergence problem can be eased in a large extent with the lowest overhead. Combined with a well-designed thread data reorganization algorithm, this solution can better mitigate GPUs' control flow divergence problem. © 2024 Copyright is held by the owner/author(s). Publication rights licensed to ACM.

关键词： Tensors

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：