检索结果-内蒙古大学图书馆

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Wang, Zheng Min, Geyong Ruan, Wenjie Department of Computer Science University of Exeter United Kingdom School of Computer Science USTC China

The implicit bias of gradient descent has long been considered the primary mechanism explaining the superior generalization of over-parameterized neural networks without overfitting, even when the training error is zero. However, the implicit bias toward adversarial robustness has rarely been considered in the research community, although it is crucial for the trustworthiness of machine learning models. To fill this gap, in this paper, we explore whether layers in neural networks collaborate to strengthen adversarial robustness during gradient descent. By quantifying this collaboration between layers using our proposed concept, co-correlation, we demonstrate a monotonically increasing trend in co-correlation, which implies a decreasing trend in adversarial robustness during gradient descent. Additionally, we observe different behaviours between narrow and wide neural networks during gradient descent. We conducted extensive experiments that verified our proposed theorems. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

IFVONet: Integrating Inter-Frame Variation and Occlusion Awareness for 3D Hand Mesh Reconstruction 10

IFVONet: Integrating Inter-Frame Variation and Occlusion Awa...

引用

10th IEEE Smart World Congress, SWC 2024

作者： Wu, Shiyao Panneerselvam, John Liu, Lu Raman, Rajeev Huang, Tianjin University of Leicester School of Computing and Mathematical Sciences Leicester United Kingdom University of Exeter Department of Computer Science Exeter United Kingdom

ISBN: (纸本)9798331520861

Reconstructing 3D hand meshes from video files is significantly challenging due to objects in the video often occluding the hand during manipulation. These occlusions can greatly reduce the quality of information extracted from the obscured regions and decrease temporal hand coherence over time. Existing approaches focus primarily on global occlusion regions but overlook temporal hand coherence, which limits their performance. Herein, we propose a novel framework called IFVONet, designed to improve 3D hand mesh reconstruction by effectively capturing inter-frame variations and improving the recovery of global occlusions. IFVONet comprises three key components: (1) Pixel-Domain Variation Module for identifying inter-frame variations across adjacent frames, enhancing temporal hand coherence. (2) Enhanced Global Occlusion Recovery Module for integrating hand information into global occlusion representation, thereby improving the accuracy of occlusion feature recovery. (3) Hand Regression Module for dynamically aggregating hand information from inter-frame variations and globally recovered occlusion features into comprehensive hand representations, ultimately leading to enhanced 3D hand reconstruction. Extensive experiments on the HO3D-v2 and HO3D-v3 datasets demonstrate that our proposed IFVONet achieves state-of-the-art performance on both 3D hand mesh reconstruction and pose estimation. © 2024 IEEE.

关键词： Mesh generation

来源：评论

学校读者我要写书评

暂无评论

Construction of hydrangea-like core-shell SiO_(2)@Ti_(3)C_(2)T_(x)@CoNi microspheres for tunable electromagnetic wave absorbers

引用

Journal of Advanced Ceramics 2023年第4期12卷 711-723页

作者： Huanhuan Niu Xuewen Jiang Yongde Xia Hailong Wang Rui Zhang Hongxia Li Bingbing Fan Yanchun Zhoua School of Material Science&Engineering Zhengzhou UniversityZhengzhou 450001China Department of Engineering University of ExeterExeter EX44SBUK School of Materials Science and Engineering Luoyang Institute of Science and TechnologyLuoyang 471023China Sinostel Luoyang Institute of Refractories Research Co. Ltd.Luoyang 471039China

Ti_(3)C_(2)T_(x)MXene shows great potential in the application as microwave absorbers due to its high attenuation ***,excessively high permittivity and self-stacking are the main obstacles that constrain its wide range of *** tackle these problems,herein,the microspheres of SiO_(2)@Ti_(3)C_(2)T_(x)@CoNi with the hydrangea-like core-shell structure were designed and prepared by a combinatorial electrostatic assembly and hydrothermal reaction *** microspheres are constructed by an outside layer of CoNi nanosheets and intermediate Ti_(3)C_(2)T_(x)MXene nanosheets wrapping on the core of modified SiO_(2),engendering both homogenous and heterogeneous *** trilayer SiO_(2)@Ti_(3)C_(2)T_(x)@CoNi microspheres are“magnetic microsize supercapacitors”that can not only induce dielectric loss and magnetic loss but also provide multilayer interfaces to enhance the interfacial *** optimized impedance matching and core-shell structure could boost the reflection loss(RL)by electromagnetic *** synthesized SiO_(2)@Ti_(3)C_(2)T_(x)@CoNi microspheres demonstrate outstanding microwave absorption(MA)performance benefited from these *** obtained RL value was-63.95 dB at an ultra-thin thickness of 1.2 mm,corresponding to an effective absorption bandwidth(EAB)of 4.56 *** work demonstrates that the trilayer core-shell structure designing strategy is highly efficient for tuning the MA performance of MXene-based microspheres.

关键词： SiO_(2)@Ti_(3)C_(2)T_(x)@CoNi hydrangea-like core-shell structure spherical capacitor high-performance microwave absorption(MA) interfacial polarization impedance match

来源：评论

学校读者我要写书评

暂无评论

A Novel Graph-Representation-Based Multi-Scale Feature Fusion Approach for Road Segmentation 21

A Novel Graph-Representation-Based Multi-Scale Feature Fusio...

引用

21st International Conference on Networking, Sensing and Control, ICNSC 2024

作者： Li, Zhenyang Song, Siyang Chen, Fuxiang Liu, Lu School of Computing and Mathematical Sciences University of Leicester Leicester United Kingdom University of Exeter Department of Computer Science Exeter United Kingdom

ISBN: (纸本)9798350365221

Exploiting contextual information is important for road segmentation tasks. In contrast to existing work primarily focusing on expanding the receptive field and stacking network layers, we propose a novel Graph-representation-based Multi-scale Feature Fusion U-net model (GMFFUnet) to address this problem. Our method enhances global context extraction by establishing connections between features of varying types and scales. Specifically, our model projects the entire feature pyramid into multiple interaction spaces and models the complex relationships among them to construct a graph and facilitate information exchange. The enhanced features are then fused back into the original feature pyramid. Our method provides substantial benefits on road segmentation tasks with complex backgrounds. Comparative experiments were carried out to evaluate the performance of the proposed method, and the results demonstrated that our proposed method outperforms all SOTA methods on the Massachusetts road dataset. © 2024 IEEE.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Impact Assessment of Memetic Operators on Communication Topologies for the PSO

Impact Assessment of Memetic Operators on Communication Topo...

引用

2023 IEEE Latin American Conference on Computational Intelligence, LA-CCI 2023

作者： Silva, Carlos Santana, Clodomir Bastos-Filho, Carmelo Department of Computer Engineering University of Pernambuco Recife Brazil Department of Computer Science University of Exeter Exeter United Kingdom

ISBN: (纸本)9798350348071

Swarm intelligence has emerged as a promising research area due to its potential to solve complex problems. The Particle Swarm Optimisation (PSO) algorithm is widely used in this area due to its efficiency and easy implementation. However, PSO performance can be affected by the choice of the communication topology, which defines the interaction between particles. In this work, we propose a comparative analysis of the impact of memetic operators on three distinct PSO topologies. We explore Interaction Networks’ use to analyse individuals’ social interaction in PSO-based algorithms with memetic operators. Furthermore, we employ the Network Portrait Divergence metric to compare the structural similarity between networks of the algorithms and the Interaction Diversity to measure the diversity of information flow within the swarm. The results obtained in this study provide insights into the topology selection for the PSO and the effect of memetic operators on them. © 2023 IEEE.

关键词： Swarm intelligence

来源：评论

学校读者我要写书评

暂无评论

Multi-Objective Optimization of the Ultrasonic Scalpel Rod and Tip with Improved Performance:Vibration Frequency,Amplitude,and Service Life

引用

Chinese Journal of Mechanical engineering 2024年第4期37卷 107-119页

作者： Jiaqi Zhao Yuhao Zhai Xuzhe Jia Naiwen Deng Kunxu Li Guangchao Han Rong Chen Dong Wang Wei Bai School of Mechanical Engineering and Electronic Information China University of GeosciencesWuhan 430074China Shenzhen Research Institute China University of GeosciencesShenzhen 518057China College of Engineering Mathematics and Physical SciencesUniversity of ExeterExeter EX44QFUK Shenzhen Huazhong University of Science and Technology Research Institute Shenzhen 518000China

Ultrasonic scalpel design for minimally invasive surgical procedures is mainly focused on optimizing cutting ***,an important issue is the low fatigue life of traditional ultrasonic scalpels,which affects their long-term reliability and effectiveness and creates hidden dangers for *** this study,a multi-objective optimal design for the cutting performance and fatigue life of ultrasonic scalpels was proposed using finite element analysis and fatigue *** optimal design parameters of resonance frequency and amplitude were *** setting the transition fillet and keeping the gain structure away from the node position to enable the scalpel to have a high service life with excellent cutting *** frequency modulation method of setting the vibration node bosses at the node position and setting the vibration antinode grooves at the antinode position was ***,the mechanism of the influence of various design elements,such as tip,shank,node position,and antinode position,on the resonance frequency,amplitude,and fatigue life of the ultrasonic scalpel was analyzed,and the optimal design principles of the ultrasonic scalpel were *** proposed ultrasonic scalpel design was confirmed by simulations,impedance measurements,and liver tissue cutting experiments,demonstrating its feasibility and enhanced *** research introduces innovative design strategies to improve the fatigue life and performance of ultrasonic scalpels to address an important issue in minimally invasive surgery.

关键词： Multi-objective optimal design Minimally invasive surgical procedures Service life Ultrasonic scalpel

来源：评论

学校读者我要写书评

暂无评论

A Knowledge Flow Empowered Cognitive Framework for Decision Making With Task-Agnostic Data Regulation

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第5期5卷 2304-2318页

作者： Huang, Liming Wu, Yulei Tempini, Niccolo Department of Computer Science Faculty of Environment University of Exeter Science and Economy ExeterEX4 4RN United Kingdom the Faculty of Science and Engineering and Bristol Digital Futures Institute University of Bristol BristolBS8 1UB United Kingdom Department of Social and Political Sciences Philosophy and Anthropology Faculty of Humanities Arts and Social Sciences University of Exeter ExeterEX4 4RN United Kingdom

The extreme complexity of many real-world tasks poses considerable challenges to agents' decision making. Most existing models only rely on task-related data for knowledge learning, while ignoring the important influence of potential task-agnostic factors. Effective learning coupled with both task-related and task-agnostic data can strongly enrich the agent's knowledge and improve its decision making performance. Furthermore, many existing learning models simply leverage data to learn knowledge but fail to express the thought process of decision making as humans do, which significantly limits their explanatory capability. To this end, we propose a novel knowledge flow empowered cognitive framework for real-world tasks. To obtain more reliable and trustworthy knowledge, a bottom-up knowledge learning model is developed, which incorporates both task-related data and task-agnostic data for comprehensive knowledge accumulation and value assessment of influencing factors. To demonstrate the thought process of decision making, a top-down knowledge expression model is proposed to coordinate different influencing factors by a knowledge flow structure. Two real-world case studies, including traffic anomaly detection and vehicle following anomaly detection, are introduced, where task-agnostic data are presented for the first time in both tasks. Experimental evaluation demonstrates the strong necessity of incorporating task-agnostic data in knowledge accumulation for real-world tasks and the effectiveness of our cognitive framework. © 2020 IEEE.

关键词： Knowledge representation

来源：评论

学校读者我要写书评

暂无评论

vPERFORM: The Development of Footfall Loading Models for Human Walking on Vibrating Surfaces 41st

vPERFORM: The Development of Footfall Loading Models for Hu...

引用

41st IMAC, A Conference and Exposition on Structural Dynamics, 2023

作者： Zhang, Sigong Živanović, Stana Williams, Genevieve School of Engineering Newcastle University Newcastle Upon Tyne United Kingdom Faculty of Environment Science and Economy University of Exeter Exeter United Kingdom Faculty of Health and Life Sciences University of Exeter Exeter United Kingdom

ISBN: (纸本)9783031366628

The design for structural vibration induced by human walking is currently based on loading models obtained on rigid surfaces. It is largely unknown how footstep forces would be modified due to human-structure interaction if the surfaces are vibrating. The 2-year project, vPERFORM, aims to identify vertical vibration conditions under which the human-structure interaction occurs and model the interaction to reflect experimental observations of human walking on a vibrating structure. Experiments were carried out in a state-of-the-art facility, VSimulators, to investigate the influence of vibration amplitude, frequency and exposure time on the interaction during normal walking on a treadmill placed on a vibrating platform. Fifteen test subjects were recruited, and ground reaction forces and kinematics data were recorded using force plates and motion capture systems, respectively. In addition, the metabolic cost was also measured to clarify the role of energy optimization in human walking in a vibrating environment. This paper presents details of the experimental program as well as selected preliminary results on ground reaction forces of human walking on the vibrating platform. © 2024, The Society for Experimental Mechanics, Inc.

关键词： Structural dynamics

来源：评论

学校读者我要写书评

暂无评论

Preference-Based Multi-Objective Optimization with Gaussian Process

Preference-Based Multi-Objective Optimization with Gaussian ...

引用

2023 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2023

作者： Huang, Tian Li, Ke University of Electronic Science and Technology of China Department of Computer Science and Engineering Chengdu China University of Exeter Department of Computer Science Exeter United Kingdom

ISBN: (纸本)9798350337020

Traditional evolutionary multi-objective optimization (EMO) algorithm is to generate a set of non-dominated solutions on the Pareto front (PF). However, this technique falls short of delivering the outcomes for multi-objective optimization problems (MOPs) containing user preference. In this paper, we present a novel EMO algorithm that incorporates user preferences via a decision maker (DM). Our approach comprises three modules: consultation, preference elicitation and optimization. The DM undertakes the consultation and preference elicitation using Gaussian process (GP) to provide preference information. We employ the decomposition-based EMO algorithm (i.e., MOEA/D) for optimization. The experiment comprises two sessions. Firstly, we simulate the decision maker module with GP. Secondly, we simulate our proposed method and compare its performance with existing interactive optimization algorithms. Our research proposes a new preference-based EMO algorithm that addresses the shortcomings of traditional techniques and unlocks new possibilities for multi-objective optimization. © 2023 IEEE.

关键词： evolutionary algorithms Gaussian process Multi-objective optimization preference learning

来源：评论

学校读者我要写书评

暂无评论

Cost-Effective Task offloading and Resource Scheduling for Mobile Edge Computing in 6G Space-Air-Ground Integrated Network

引用

IEEE Internet of Things Journal 2025年第12期12卷 19428-19442页

作者： Zhu, Wenwu Deng, Xiaoheng Gui, Jinsong Zhang, Honggang Min, Geyong The School of Electronic Information Central South University Changsha410075 China The Shenzhen Research Institue Central South University China The Engineering Department Umass Boston BostonMA02125 United States The Department of Computer Science University of Exeter ExeterEX4 4QF United Kingdom

With the advent of the sixth-generation (6G) wireless communications, transmission speeds are projected to exceed tenfold those of 5G, reaching theoretical peak download speeds of up to 1 Tbps. Data transmission capacity and speed will be significantly enhanced, enabling emerging applications such as mixed reality, federated learning, and digital twins, driving exponential data traffic growth. To address this, the space-air-ground integrated network (SAGIN) combines satellite, aerial, and ground communication technologies, offering seamless global coverage and high-speed connectivity. In this paper, we proposes a SAGIN framework integrated with mobile edge computing (MEC) to jointly optimize system energy consumption and delay costs. Specifically, we decompose the optimization problem into three sub-problems: 1) Unmanned Aerial Vehicle (UAV) computational resource allocation;2) satellite computational resource allocation;and 3) task offloading and channel allocation. The subproblems are then transformed and addressed using Newton’s interior point method and the deep reinforcement learning DQN algorithm to derive optimal allocation strategies for UAV and satellite computing resources, along with task offloading and channel resources, that our proposed algorithm effectively reduces system energy consumption and delay costs compared to other algorithms. © 2014 IEEE.

关键词： Unmanned aerial vehicles (UAV)

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：