检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Song, Juan He, Jiaxiang Feng, Mingtao Wang, Keyan Li, Yunsong Mian, Ajmal Xidian University Xi’an710071 China The Department of Computer Science and Software Engineering The University of Western Australia Perth CrawleyWA6009 Australia

Diffusion probabilistic models have recently achieved remarkable success in generating high-quality images. However, balancing high perceptual quality and low distortion remains challenging in image compression applications. To address this, we propose an efficient Uncertainty-Guided image compression approach with wavelet Diffusion (UGDiff). Our approach focuses on high frequency compression via the wavelet transform, since high frequency components are crucial for reconstructing image details. We introduce a wavelet conditional diffusion model for high frequency prediction, followed by a residual codec that compresses and transmits prediction residuals to the decoder. This diffusion prediction-then-residual compression paradigm effectively addresses the low fidelity issue common in direct reconstructions by existing diffusion models. Considering the uncertainty from the random sampling of the diffusion model, we further design an uncertainty-weighted rate-distortion (R-D) loss tailored for residual compression, providing a more rational tradeoff between rate and distortion. Comprehensive experiments on two benchmark datasets validate the effectiveness of UGDiff, surpassing state-of-the-art image compression methods in R-D performance, perceptual quality, subjective quality, and inference time. Our code is available at: https://***/hejiaxiang1/WaveletDiffusion/tree/main. Copyright © 2024, The Authors. All rights reserved.

关键词： Wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

Schrödinger Mechanisms: Optimal Differential Privacy Mechanisms for Small Sensitivity

Schrödinger Mechanisms: Optimal Differential Privacy Mechan...

引用

IEEE International Symposium on Information Theory

作者： Wael Alghamdi Shahab Asoodeh Flavio P. Calmon Juan Felipe Gomez Oliver Kosut Lalitha Sankar School of Engineering and Applied Sciences Harvard University Department of Computing and Software McMaster University School of Electrical Computer and Energy Engineering Arizona State University

We consider the problem of designing optimal differential privacy mechanisms with a favorable privacy-utility tradeoff in the limit of a large number n of compositions (i.e., sequential queries). Here, utility is measured by the average distance between the mechanism's input and output, evaluated by a cost function c. We show that if n is sufficiently large and the sensitivities of all queries are small, then the optimal additive noise mechanism has probability density function fully characterized by the ground-state eigenfunction of the Schrödinger operator with potential c. This leads to a family of optimal mechanisms, dubbed the Schrödinger mechanisms, depending on the choice of the cost function. Instantiating this result, we demonstrate that for c(x) = x 2 the Gaussian mechanism is optimal, and for c(x) = |x|, the optimal mechanism is obtained by the Airy function, thereby leading to the Airy mechanism.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Human-Cyber-Physical Automata and Their Synthesis 1

引用

19th International Colloquium on Theoretical Aspects of Computing, ICTAC 2022

作者： Zhang, Miaomiao Liu, Wanwei Tang, Xiaochen Du, Bowen Liu, Zhiming School of Software Engineering Tongji University Shanghai China College of Computer Science National University of Defense Technology Changsha China Department of Computer Science University of Warwick Coventry United Kingdom School of Computer and Information Science Southwest University Chongqing China School of Software Northwest Polytechnical University Xi’an China

A cyber-physical system (CPS) is now well understood as an integration of hierarchical composition digital computing (or information processing) systems and physical systems [3, 10, 14]. © 2022, The Author(s), un... 详细信息

ISBN: (数字)9783031177156

ISBN: (纸本)9783031177149

关键词： Cyber Physical System

来源：评论

学校读者我要写书评

暂无评论

Identify the potential pathways and candidate biomarkers of stroke associated with bipolar disorder: Bioinformatics and system biology approach

引用

Journal of Proteins and Proteomics 2025年第2期16卷 141-159页

作者： Hosen, Md. Faruk Basar, Md. Abul Yasmin, Mst. Farjana Hasan, Md. Rakibul Uddin, Muhammad Shahin Department of Information and Communication Technology (ICT) Mawlana Bhashani Science and Technology University (MBSTU) Tangail Bangladesh Department of Computing and Information System (CIS) Daffodil International University (DIU) Dhaka Bangladesh Department of Computer Science and Engineering (CSE) Uttara University Dhaka Bangladesh Department of Software Engineering (SWE) Daffodil International University (DIU) Dhaka Bangladesh

The field of computational bioinformatics and systems biology analysis is rapidly expanding due to advancements in bioinformatics tools. Bipolar disorder (BD) is a severe psychiatric condition affecting both adults and adolescents. In recent years, the risk of stroke has increased among individuals with BD. Background studies indicate that BD and stroke share numerous biochemical and genetic characteristics. Due to traditional endocrinological approaches, no collaborative research exists, and no treatment options have been established for BD and stroke patients. The aim of this study is to identify common molecular pathways and potential therapeutic targets in BD and stroke that could be utilized to predict disease progression. To achieve this, shared genes were identified to determine common pathways. Based on the biochemical, molecular, and genetic interactions among these shared genes, this study identifies the most significant hub genes. To further explore these interactions, analysis such as the protein–protein interaction (PPI) network, topological properties, enrichment analysis, co-expression network, gene regulatory network (GRN), and physical interaction network are conducted. The GO annotation indicates that most of the genes are linked to MAP kinase activity and tumor necrosis factor receptor superfamily binding. The KEGG pathway analysis revealed that the predominant genes are associated with apoptosis and the IL-18 signaling pathway. This comparison helps understand the biochemical and genetic characteristics shared between BD and stroke. Predictive drug analysis has identified consistent potential biomarkers associated with both conditions, providing a scientific foundation for investigating their diagnosis, treatment, and prognosis. Finally, chemical experiments could be conducted to further validate the efficacy of these drugs.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Enhancing Deep Convolutional Neural Networks to Detect Pedestrians Effectively in Various Urban Environments

Enhancing Deep Convolutional Neural Networks to Detect Pedes...

引用

International Conference on Electrical engineering and Information & Communication Technology (ICEEICT)

作者： Md. Mehedi Hasan Nakib Aman Md. Wadud Jahan Software Engineer I Developer eXperience Hub Rajshahi Bangladesh Department of Computer Science and Engineering Pabna University of Science and Technology Pabna Bangladesh

ISBN: (数字)9798350385779

ISBN: (纸本)9798350385786

In computer vision, accurately identifying pedestrians is a crucial difficulty, especially in applications where safety is of the highest priority, such autonomous driving and urban planning. Deep learning has made progress, but challenges such changing conditions, barriers, and processing power requirements still exist. In this paper, we present an improved deep convolutional neural network (DCNN) designed with pedestrian identification in mind. Improving accuracy in various settings while maximizing processing efficiency is our goal. We carefully prepare the data and create a novel DCNN architecture that can handle a variety of pedestrian appearances and urban surroundings by utilizing well-known datasets such as Penn-Fudan and PnPLO. Our model performs exceptionally well in terms of accuracy, speed, and durability, according to a thorough evaluation that compares it to recognized benchmarks. In addition to making important advances, our study has shown novel approaches to enhancing performance in challenging environments, such as the analysis of transfer learning and the deployment of real-time systems. This work paves the way for future developments in intelligent, self-governing systems and is a maj or leap forward in pedestrian detection technology.

关键词： Training Pedestrians Computational modeling Surveillance Transfer learning Urban planning computer architecture

来源：评论

学校读者我要写书评

暂无评论

Temporal Sensitivity Analysis of Internal Temperature Informed Charging Algorithms and Rapid Thermal Management System for E-mobility

Temporal Sensitivity Analysis of Internal Temperature Inform...

引用

IEEE Energy Conversion Congress and Exposition, ECCE

作者： Akash Samanta Chandan Chetri Dominic Karnehm Antje Neve Sheldon Williamson Department of Electrical Computer and Software Engineering Ontario Tech University Oshawa Canada Institute of Intelligent Distributed Systems Bundeswehr University Munich Germany

ISBN: (数字)9798350376067

ISBN: (纸本)9798350376074

The experimental studies presented in this paper reveal that existing thermal management systems (TMS) and temperature-informed charging algorithms exhibit a response time lag of at least 5.3 minutes due to their reliance on surface temperature measurements. The results indicate that changes in the internal thermal state of lithium-ion batteries (LIBs), induced by variations in charging currents, take an average of 2 minutes to manifest on the battery surface, particularly evident in cylindrical cells. Current thermal management systems for automotive battery packs solely rely on surface temperature measurements, neglecting the approximately 5.8°C temperature difference between the core and surface in TMS control. Consequently, changes in the battery's thermal state due to internal heat losses are not promptly detected by surface-mounted temperature sensors. This delayed response time accelerates battery degradation and increases the risk of thermal runaway events. In this study, temperature-informed fast charging algorithms, tested under various ambient conditions for LIBs, along with a comparative analysis, demonstrate that response time can be reduced by at least 2 minutes by considering internal temperature rather than relying solely on surface temperature measurements. Moreover, accounting for the temperature difference between the core and surface facilitates rapid TMS control and health-conscious fast charging, thereby mitigating the risk of thermal runaway events.

关键词： Temperature measurement Temperature sensors Heating systems Sensitivity analysis Surface charging Thermal management Time measurement Fast charging Time factors Thermal degradation

来源：评论

学校读者我要写书评

暂无评论

Fdc: Feature Dropout Consistency for Unsupervised Domain Adaptation Semantic Segmentation

SSRN

引用

SSRN 2024年

作者： Yao, Chaoyu Fan, Wanshu Wang, Cong Yang, Xin Wei, Xiaopeng Zhou, Dongsheng School of Computer Science and Technology Dalian University of Technology China National and Local Joint Engineering Laboratory of Computer Aided Design School of Software Engineering Dalian University China Department of Computing The Hong Kong Polytechnic University China

In Unsupervised Domain Adaptation Semantic Segmentation (UDASS), while self-training techniques have become one of the most effective methods to date, the absence of target labels makes models susceptible to overfitting. To address this problem, consistency techniques regularize the model by perturbing the target domain and enforcing consistent pixel predictions before and after the perturbation. Consistency techniques are based on the idea that a well-performing model should yield consistent predictions for the target data stream before and after perturbations. In this paper, we introduce the Feature Dropout Consistency (FDC) module, a novel approach that instigates perturbations at the feature level. This is achieved by implementing random feature dropout between the encoder and decoder of the student network. Subsequently, a feature dropout consistency loss is applied to minimize discrepancies in predictions between the perturbed student network and its teacher counterpart. Notably, FDC seamlessly integrates with existing self-training methodologies and, when paired with input perturbation, delves into an even broader perturbation spectrum. Through rigorous experimentation in standard UDA settings—training with synthetic labeled and real unlabeled data—FDC consistently outperforms baseline models, establishing new benchmarks on GTAV →Cityscapes and SYNTHIA →Cityscapes. © 2024, The Authors. All rights reserved.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Comparative analysis of various models for image classification on Cifar-100 dataset

Comparative analysis of various models for image classificat...

引用

作者： YuYu Zheng HaoXuan Huang JunMing Chen School of Optoelectronic Information and Computer Engineering University of Shanghai for Science and Technology Department of Computing and Software McMaster University Department of Mathematics University of CaliforniaLos Angeles

Nowadays,people developed various convolutional neural network(CNN) based models for computer *** famous models,such as GoogLeNet,Residual Network(ResNet),Visual Geometry Group(VGG),and You Only Look Once(YOLO),have different architecture and *** which model to use may be a troublesome problem for those just starting to study image *** solve this problem,we introduce the GoogLeNet,ResNet-18,and VGG-16 models,comparing their architecture,features,and *** we give our suggestions based on the test results to help beginners choose a suitable *** conducted experiments to train and test GoogLeNet,ResNet-18,and VGG-16 on the Cifar-100 datasets with the same *** on the test results(test accuracy,average test loss,training loss),we analyze the figures for trends,key points,increase rate,and other *** we combine the architecture of each model to make our *** experimental results show that ResNet-18 can be a good choice when training the model with the Cifar-100 datasets because it performs well after training and has a low time *** Net-18 also has the fastest convergence *** would be the second choice because it functions similarly to ResNet-18 and is even ***,training GoogLeNet is a time-consuming *** is not recommended in this experiment because it has the worst performance and similar training complexity compared with ResNet-18.

关键词： GoogLeNet ResNet VGG Image Classification Architecture Comparison

来源：评论

学校读者我要写书评

暂无评论

Computational Model of Swarm Algorithm for Optimizing Process of Keywords Extraction from Text Information Presented as Graph

Computational Model of Swarm Algorithm for Optimizing Proces...

引用

International Conference on Information Technologies in engineering Education (Inforino)

作者： Yury Kravchenko Elmar Kuliev Daniil Kravchenko Sergey Rodzin Nikolay Pavlov Lada Rodzina Dept. of Computer Aided Design Southern Federal University Taganrog Russia Department of Software Engineering Southern Federal University Taganrog Russia

ISBN: (数字)9798350371369

ISBN: (纸本)9798350371376

The article describes the scientific problem of optimizing the processes of acquiring knowledge when processing textual information presented in the form of a graph, based on the development of a modified bioinspired algorithm for extracting keywords. The solution proposed by the authors differs from the known ones by introducing new attractive operators and behavioral patterns of bacterial agents into the computational model of the bacterial swarm algorithm. This approach allows you to balance the ratio of the speed of convergence of the algorithm with the diversification of the search space for solutions. The article also presents an analytical study of the features of two types of parsing: constituency parsing and dependency parsing. As part of the study, algorithms for syntactic parsing of dependencies between elements of text information and a graph-based dependency parsing algorithm are described. The formulation of the problem of keyword extraction is given. The task of achieving a balance between accuracy and speed of a linguistic parser is an open topic that requires further research due to the importance of this problem for improving the efficiency of text analysis, especially in applications that require real-time accuracy. To increase the efficiency of solving the problem, the authors propose a modified bioinspired algorithm for extracting keywords. The computational experiment carried out allows us to conclude that the modified bacterial optimization method proposed by the authors is in a leading position, which confirms the need to continue research into the applicability of bioinspired methods for solving problems of acquiring knowledge when processing text information presented in the form of a graph. This article is focused on the direction of study in the master’s program “Informatics and computer Science”, as well as the scientific specialty of postgraduate training “Artificial Intelligence and Machine Learning”.

关键词： Microorganisms Accuracy Text analysis Machine learning algorithms Computational modeling Optimization methods Syntactics

来源：评论

学校读者我要写书评

暂无评论

Downlink Transmit-End Beam Refinement for mmWave Communications Under Mobility

Downlink Transmit-End Beam Refinement for mmWave Communicati...

引用

2023 International Conference on Electrical engineering and Photonics, EExPolytech 2023

作者： Fokin, Grigoriy Grishin, Ilya Okuneva, Darina The Bonch-Bruevich Saint Petersburg State University of Telecommunications Laboratory of Software Defined Radio Saint Petersburg Russia The Bonch-Bruevich Saint Petersburg State University of Telecommunications Department of Communication Networks and Data Transmission Saint Petersburg Russia The Bonch-Bruevich Saint Petersburg State University of Telecommunications Department of Software Engineering and Computer Facilities Saint Petersburg Russia

ISBN: (纸本)9798350310481

This article discusses a method for adjusting the radiation pattern of a base station antenna array when conducting radio communications in the millimeter wavelength range for mobile user devices. The article describes a technique for selecting a training sequence of beams that minimizes the Cramer-Rao lower bound when estimating the transmission direction. The proposed technique makes it possible to reduce the beam adjustment time for transmission by 16 times compared to the conventional protocol, which improves the quality of service for highly mobile users. © 2023 IEEE.

关键词： Beamforming

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：