检索结果-内蒙古大学图书馆

Graph Convolutional Network Robustness Verification Algorithm Based on Dual Approximation 1

25th International Conference on Formal Engineering methods, ICFEM 2024

作者： An, Dongdong Zhang, Hao Zhao, Qin Liu, Jing Shi, Jianqi Huang, Yanhong Yang, Yang Liu, Xu Qin, Shengchao Shanghai Engineering Research Center of Intelligent Education and Bigdata Shanghai Normal University Shanghai200234 China Shanghai Key Laboratory of Trustworthy Computing East China Normal University Shanghai200062 China Shanghai Formal-Tech Information Technology Co. LTD. Shanghai China Guangzhou Institute of Technolog Xidian University Shaanxi China ICCT and ISN Laboratory Xidian University Shaanxi China

ISBN: (数字)9789819606177

ISBN: (纸本)9789819606160

With the continuous development of Graph Neural Network (GNN) technologies, securing their robustness is crucial for their broad adoption in practical applications. Although various verification methods for training GNNs have been proposed, studies indicate that Graph Convolutional Networks (GCNs) remain vulnerable to adversarial attacks affecting both graph structure and node attributes. We propose a novel approach to verify the robustness of GCNs against perturbations in node attributes by employing a dual approximation technique to convexify nonlinear activation functions. This transformation changes the original non-convex problem into a more manageable convex forms. We start by applying linear relaxation to convert fixed-value features in each GCN layer into variables suitable for optimization. Next, we reframe the task of identifying the worst-case margin for a graph as a linear problem, which we solve using linear programming techniques. Given the discrete nature of graph data, we define a perturbation space that extends the data domain from discrete to continuous values. To improve the accuracy of the convex relaxation, we use a dual approximation algorithm to set bounds on the optimizable variables. Our method certifies the robustness of nodes against perturbations within a specified range and significantly improves verification accuracy compared to previous approaches. This method surpasses previous work in verification accuracy and is distinctively tailored to address the S-curve, an aspect less explored in prior research. Experimental results show that our method significantly refines the precision of robustness verification for GCNs. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Dirichlet Process-based Prediction of Gasholder Level with Self-learning Scheme in Steel Enterprises

Dirichlet Process-based Prediction of Gasholder Level with S...

引用

2022 Chinese Automation Congress, CAC 2022

作者： Zhang, Yu Chen, Long Han, Zhongyang Zhao, Jun Wang, Wei Key Laboratory of Intelligent Control and Optimization for Industrial Equipment Ministry of Education Dalian China

ISBN: (纸本)9781665465335

Gasholder level prediction is very important for the workers to perform energy scheduling in steel enterprises. The gasholder level model is time-varying due to environmental disturbance. Since the traditional gasholder level prediction are often based on off-line model without the consideration of the change of working conditions, the prediction results may be inaccurate. This paper proposes a gasholder level prediction method with self-learning based on Dirichlet process, to deal with the multi-conditions structure of gas system. The Dirichlet process is introduced as a priori distribution so as to automatically decide the number of feature clusters, and then realize the automatic clustering algorithm for multiple working conditions of gasholder. Compared with other state of the art prediction methods, the advantage of this method is that the number of clusters need not to be specified in advance. The experimental results show that the method in this paper achieves better performance in the accuracy of gasholder prediction. © 2022 IEEE.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Diabetes mellitus and human immunodeficiency virus (HIV) infection in people with tuberculosis in Odisha, India

引用

Indian Journal of Tuberculosis 2024年第2期71卷 147-152页

作者： Giri, Sidhartha Sahu, Priyanka Kanungo, Srikanta Bal, Himadri Bhusan Kumar, Sujeet Kar, Sarita Mohanty, Triyambakesh Turuk, Jyotirmayee Das, Dasarathi Hota, Prasanta Kumar Pati, Sanghamitra ICMR-Regional Medical Research Centre Bhubaneswar-751023 Odisha India National Reference Laboratory for TB ICMR-Regional Medical Research Centre Bhubaneswar-751023 Odisha India Government of Odisha India

Background: modelling studies have indicated that approximately 20% of all tuberculosis (TB) cases may suffer from diabetes mellitus (DM). DM increases the risk of developing active TB disease by 2–3 times. People living with HIV (PLHIV) are more likely to develop TB disease, and TB is a leading cause of hospitalization and death among PLHIV. Despite the substantial burden of DM and HIV in India, few studies have evaluated the prevalence of DM and HIV among active cases of TB, and its impact on the treatment outcome for TB. This study evaluated the burden of HIV and DM in TB cases from Odisha during 2019, and its impact on the TB treatment outcome. methods: The study utilized data on TB patients of Odisha during 2019, from the NIKSHAY portal, the health management information system (HMIS) of TB in India. This is a retrospective observational registry-based cohort study, which evaluated a linkage between socio-demographic predictors, clinical diagnostic and treatment predictors, time of treatment predictors, and co-morbidity with TB. Data were retrieved electronically in Microsoft-Excel and analysis was done using STATA 16 (StataCorp. 2019, College Station, TX: StataCorp LLC). Results: Data for 47,831 TB cases of Odisha as study population was extracted from the Nikshay application for the year 2019. The highest prevalence (31.1%, 14,863/47,831) of TB was observed among young participants aged 15–30 years, whereas the prevalence was least among children <14 years (4.4%, 2124/47,831). Males had a higher prevalence of TB (66.7%, 31,878/47,831). Of the 47,831 TB cases included in the study, 7.6% (3659/47,831) had diabetes mellitus (DM), along with TB. 1.2% (571/47,831) had HIV along with TB, while only 0.08% (37/47,831) had both DM and HIV along with TB. 88.2% (3148/3569) of cases with DM and TB had a favorable outcome, compared to 82.3% (449/541) of cases with HIV and TB. People with TB who did not have DM had a significantly higher favorable outcome (OR 1.6, 95% CI 1

关键词： Diabetes mellitus Human immunodeficiency virus Odisha Tuberculosis

来源：评论

学校读者我要写书评

暂无评论

A Constrained Multi-objective optimization for the Evaluation and Selection of Monitored Bridges in Road Network

A Constrained Multi-objective Optimization for the Evaluatio...

引用

Hydraulic and Civil Engineering & Smart Water Conservancy and Intelligent Disaster Reduction Forum (ICHCE & SWIDR), International Conference on

作者： Fangyuan Xu Sugong Chao Yuanli Chen Jiaying Wang Key Laboratory for Detection and Maintenance Technology of Road and Bridge of Zhejiang Province Zhejiang Institute of Communications Hangzhou China

ISBN: (纸本)9781665465540

The selection of bridges to be monitored in an extensive road network requires comprehensive considerations such as total monitoring cost, bridge types and span, traffic flow, regional economic radiation and bridge service performance, and might control the total quantity of monitoring bridges within a certain proportion. Based on evolutionary algorithm NSGA2, this paper improved the mutation algorithm with respect to the binary integer (0,1) problem and adopted Tabu list to reduce the redundancy of iterative search. It proposed a constrained multi-objective optimization by introducing sparse random matrix and constraint violation indicator to restrain the proportion. Taking the selection of the monitored ones from total 207 bridges in a county road network as an example, the proposed method were used to analyze the distribution of Pareto solution set, the satisfaction of constraints and the calculation efficiency under different corresponding evolutionary simulation parameters. The result shows that the evaluation based on the crowding distance in the traditional NSGA2 cannot select the optimal bridges to be monitored effectively. By introducing a fuzzy membership degree to obtain the multi-objective decision weights, the extreme value of decision function can be used for selecting the final decision plan of monitored bridges in road network.

关键词： Bridges Costs Roads Redundancy optimization methods Hydraulic systems Search problems

来源：评论

学校读者我要写书评

暂无评论

modelling And Control Allocation Of A New VTOL Configuration

Modelling And Control Allocation Of A New VTOL Configuration

引用

2022 CSAA/IET International Conference on Aircraft Utility systems, AUS 2022

作者： Liu, Yaoxian Jiao, Zongxia Zhao, Longfei School of Automation Science and Electrical Engineering Beihang University Beijing China Ningbo Institute of Technology Beihang University Ningbo China Science and Technology on Aircraft Control Laboratory Beihang University Beijing China Research Institute for Frontier Science Beihang University Beijing China

ISBN: (数字)9781839537769

ISBN: (纸本)9781839537769

In this paper, a new tandem wing distributed propulsion vertical takeoff and landing configuration is investigated, and the modeling and control assignment of this new configuration is described. Six-degree-of-freedom full-state dynamics modeling and simulation modeling are completed first. A suitable attitude representation is chosen and the analysis starts with the flight characteristics of the tandem-wing vehicle. Based on the flight principle and its force and moment generation mechanism, the dynamics and kinematic equations were studied, and a nonlinear mathematical model was constructed using a suitable method, which laid the foundation for the subsequent control system design verification. Then, on the basis of the constructed model, the derivation of the control efficiency matrix and the study of the control allocation method were completed. The study of control allocation includes the linearization of control actuators, the design of control allocators for transition states and the study and comparative analysis of three allocation strategies: pseudo-inverse, weighted pseudo-inverse and direct allocation methods. © 2022 IET Conference Proceedings. All rights reserved.

关键词： Nonlinear equations

来源：评论

学校读者我要写书评

暂无评论

A Mapping-based Dynamic Semi-Online Task Scheduling Method for Minimizing Energy in Edge Computing 23

A Mapping-based Dynamic Semi-Online Task Scheduling Method f...

引用

23rd IEEE International Conference on High Performance Computing and Communications, 7th IEEE International Conference on Data Science and systems, 19th IEEE International Conference on Smart City and 7th IEEE International Conference on Dependability in Sensor, Cloud and Big Data systems and Applications, HPCC-DSS-SmartCity-DependSys 2021

作者： Zhao, Hui Feng, Nanzhi Meng, Fanxin Wang, Quan Wan, Bo Wang, Jing School of Computer Science and Technology Xidian University Xi'an China Key Laboratory of Smart Human-Computer Interaction and Wearable Technology of Shaanxi Province Xi'an China

ISBN: (纸本)9781665494571

Semi-online task scheduling in the edge computing platform refers to a scheduling scenario where there exists unknown performance computing nodes in the system. Existing online and offline task scheduling methods may lead to long makespan or transmission time due to the influence of unknown nodes, and existing semi-online methods are not fully applicable to the scenario discussed in this paper, which aggravates the problem of high energy consumption in edge computing plat-form. To solve this problem, this paper proposes a Mapping-based Dynamic Semi-online Scheduling(MDSS) strategy for edge computing platform. Firstly, we consider three main factors that affect the energy consumption in edge computing platform, that is, the energy consumption of task execution, task transmission and idle state. Then an energy consumption optimization problem is established from the perspectives of the processing speed, routing delay and queuing delay of the edge nodes, which correspond to those three factors. Secondly, for the unknown node, this paper introduces a mapping mechanism, which maps the performance of unknown nodes to a certain given node. Then the mapping relations are dynamically adjusted by continuously checking the task queue lengths of the mapped two-sides. Therefore, the prior knowledge of known nodes can be fully used. Thirdly, an im-proved Power-of- D algorithm based on the mapping mechanism is proposed to solving the formulated problem. It utilizes the prior knowledge of known node performance, which reduces task processing time and transmission time, so that the energy consumption can be reduced. Finally, some experiments are conducted to evaluate MDSS. Results show that the dynamic mapping mechanism is effective in the semi-online scenario and MDSS can greatly reduce the energy consumption in the edge computing platform. © 2021 IEEE.

关键词： Energy consumption Processor scheduling Heuristic algorithms Dynamic scheduling Routing Delays Task analysis

来源：评论

学校读者我要写书评

暂无评论

Research on Common Fault Diagnosis and Classification Method of Centrifugal Pump Based on ReliefF and SVM

引用

International Journal of Fluid Machinery and systems 2022年第2期15卷 287-296页

作者： Xiao, Xingxin Chen, Hui Dong, Liang Liu, Houlin Fan, Chuanhan Fluid Machinery Engineering Technology Research Center Jiangsu University Zhenjiang212003 China Science and Technology Laboratory on Liquid Rocket Engine Xi’an710100 China Xi'an Aerospace Propulsion Institute Xi’an710100 China

The centrifugal pump is an important rotating machine and it is very critical to identify and differentiate among its common faults as quickly and accurately as possible. Based on the ReliefF algorithm and the sparrow search algorithm (SSA) in conjunction with support vector machine (SVM), an approach for faults classification and diagnosis of centrifugal pumps is proposed, its advantages over traditional fault diagnosis methods include a reduction in the number of characteristic parameters, shorter diagnosis times, as well as improved classification accuracy and robustness. We collected the fault data by designing a centrifugal pump fault test bench that recorded vibration signals for the rotor misalignment fault, the rotor unbalance fault, the seal ring wear fault, and normal operating conditions, and preprocessed the collected signals with Kalman filtering to remove noise interference, the time domain characteristic indexes and the frequency domain characteristic indexes of the filtered signal were extracted, each feature index is given a distinct weight using the ReliefF method, and the eigenvalues with weights less than the threshold are deleted, and the feature indexes that remain create a defect feature matrix. Particle swarm optimization (PSO), genetic algorithm (GA), and simulated annealing algorithm (SA) were used to optimize the SVM for comparison in order to verify the SSA-SVM model's performance for fault diagnosis. The comparison results show that the model has high recognition accuracy, short Classification time, and strong robustness. © 2022, Turbomachinery Society of Japan. All rights reserved.

关键词： Failure analysis

来源：评论

学校读者我要写书评

暂无评论

Atomic defects,functional groups and properties in MXenes

引用

Chinese Chemical Letters 2021年第1期32卷 339-344页

作者： Wenjun Cui Zhi-Yi Hu Raymond RUnocic Gustaaf Van Tendeloo Xiahan Sang State Key Laboratory of Advanced Technology for Materials Synthesis and Processing Wuhan University of TechnologyWuhan 430070China NRC(Nanostructure Research Centre) Wuhan University of TechnologyWuhan 430070China Center for Nanophase Materials Sciences Oak Ridge National LaboratoryOak RidgeTN 37831United States EMAT(Electron Microscopy for Materials Science) University of AntwerpAntwerpB-2020Belgium

MXenes,a new family of functional two-dimensional(2 D) materials,have shown great potential for an extensive variety of applications within the last *** defects and functional groups in MXenes are known to have a tremendous influence on the functional *** this review,we focus on recent progress in the characterization of atomic defects and functional group chemistry in MXenes,and how to control them to directly influence various properties(e.g.,electron transport,Li^(+) adsorption,hydrogen evolution reaction(HER) activity,and magnetism) of 2 D MXenes *** structural transformations such as oxidation and growth induced by atomic defects in MXenes are also *** review thus provides perspectives on property optimization through atomic defect engineering,and bottom-up synthesis methods based on defect-assisted homoepitaxial growth of MXenes.

关键词： MXenes Atomic defects Surface functional groups Electron microscopy Functional properties

来源：评论

学校读者我要写书评

暂无评论

V2I-Calib: A Novel Calibration Approach for Collaborative Vehicle and Infrastructure LiDAR systems

arXiv

引用

arXiv 2024年

作者： Qu, Qianxin Xiong, Yijin Zhang, Guipeng Wu, Xin Gao, Xiaohan Gao, Xin Li, Hanyu Guo, Shichun Zhang, Guoying The School of Artificial Intelligence China University of Mining and Technology -Beijing Beijing100083 China The State Key Laboratory of Automotive Safety and Energy School of Vehicle and Mobility Tsinghua University Beijing100084 China Computer Science and Technology in Institute of Computing Technology Chinese Academy of Sciences Beijing China

Cooperative LiDAR systems integrating vehicles and road infrastructure, termed V2I calibration, exhibit substantial potential, yet their deployment encounters numerous challenges. A pivotal aspect of ensuring data accuracy and consistency across such systems involves the calibration of LiDAR units across heterogeneous vehicular and infrastructural endpoints. This necessitates the development of calibration methods that are both real-time and robust, particularly those that can ensure robust performance in urban canyon scenarios without relying on initial positioning values. Accordingly, this paper introduces a novel approach to V2I calibration, leveraging spatial association information among perceived objects. Central to this method is the innovative Overall Intersection over Union (oIoU) metric, which quantifies the correlation between targets identified by vehicle and infrastructure systems, thereby facilitating the real-time monitoring of calibration results. Our approach involves identifying common targets within the perception results of vehicle and infrastructure LiDAR systems through the construction of an affinity matrix. These common targets then form the basis for the calculation and optimization of extrinsic parameters. Comparative and ablation studies conducted using the DAIR-V2X dataset substantiate the superiority of our approach. For further insights and resources, our project repository is accessible at https://***/MassimoQu/v2i-calib. Copyright © 2024, The Authors. All rights reserved.

关键词： Calibration

来源：评论

学校读者我要写书评

暂无评论

Dataset Distillation with Neural Characteristic Function: A Minmax Perspective

arXiv

引用

arXiv 2025年

作者： Wang, Shaobo Yang, Yicun Liu, Zhiyuan Sun, Chenghao Hu, Xuming He, Conghui Zhang, Linfeng School of Artificial Intelligence Shanghai Jiao Tong University China EPIC Lab Shanghai Jiao Tong University China Hong Kong University of Science and Technology Guangzhou China Shanghai Artificial Intelligence Laboratory China

Dataset distillation has emerged as a powerful approach for reducing data requirements in deep learning. Among various methods, distribution matching-based approaches stand out for their balance of computational efficiency and strong performance. However, existing distance metrics used in distribution matching often fail to accurately capture distributional differences, leading to unreliable measures of discrepancy. In this paper, we reformulate dataset distillation as a minmax optimization problem and introduce Neural Characteristic Function Discrepancy (NCFD), a comprehensive and theoretically grounded metric for measuring distributional differences. NCFD leverages the Characteristic Function (CF) to encapsulate full distributional information, employing a neural network to optimize the sampling strategy for the CF’s frequency arguments, thereby maximizing the discrepancy to enhance distance estimation. Simultaneously, we minimize the difference between real and synthetic data under this optimized NCFD measure. Our approach, termed Neural Characteristic Function Matching (NCFM), inherently aligns the phase and amplitude of neural features in the complex plane for both real and synthetic data, achieving a balance between realism and diversity in synthetic samples. Experiments demonstrate that our method achieves significant performance gains over state-of-the-art methods on both low- and high-resolution datasets. Notably, we achieve a 20.5% accuracy boost on ImageSquawk. Our method also reduces GPU memory usage by over 300× and achieves 20× faster processing speeds compared to state-of-the-art methods. To the best of our knowledge, this is the first work to achieve lossless compression of CIFAR-100 on a single NVIDIA 2080 Ti GPU using only 2.3 GB of memory. © 2025, CC BY.

关键词： Data assimilation

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：