检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Huang, Haitong Liu, Cheng Liu, Bo Xue, Xinghua Li, Huawei Li, Xiaowei State Key Lab of Processors Institute of Computing Technology Chinese Academy of Sciences Beijing100190 China University of Chinese Academy of Sciences Beijing100190 China Beijing Institite of Control Engineering Beijing100190 China

To ensure resilient neural network processing on even unreliable hardware, comprehensive reliability analysis against various hardware faults is generally required before the deep neural network models are deployed, and efficient error injection tools are highly demanded. However, most existing fault injection tools remain rather limited to basic fault injection to neurons and fail to provide fine-grained vulnerability analysis capability. In addition, many of the fault injection tools still need to change the neural network models and make the fault injection closely coupled with normal neural network processing, which further complicates the use of the fault injection tools and slows down the fault simulation. Due to the various fault injection implementations and error metrics, it makes the comparison between different fault-tolerant study difficult. To this end, we propose MRFI, a highly configurable multi-resolution fault injection tool for deep neural networks. It enables users to modify an independent fault configuration file rather than neural network models for the fault injection and vulnerability analysis. Particularly, it integrates extensive fault analysis functionalities from different perspectives and enables multi-resolution investigation of the vulnerability of neural networks. In addition, it does not modify the major neural network computing framework of PyTorch. Hence, it allows parallel processing on GPUs naturally and exhibits fast fault simulation according to our experiments. Moreover, we also have the fault injection calibrated with fault simulation with architectural details and validate the accuracy of the proposed fault injection. Finally, MRFI is also open sourced on Github1 © 2023, CC BY.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

Multi-Target Tracking Using a Swarm of UAVs by Q-learning Algorithm

Multi-Target Tracking Using a Swarm of UAVs by Q-learning Al...

引用

Sensor Signal Processing for Defence (SSPD)

作者： Seyed Ahmad Soleymani Shidrokh Goudarzi Xingchi Liu Lyudmila Mihaylova Wenwu Wang Pei Xiao Centre for Vision Speech and Signal Processing (CVSSP) The University of Surrey UK School of Computing and Engineering University of West London UK Department of Automatic Control and Systems Engineering The University of Sheffield UK Institute for Communication Systems (5GIC) The University of Surrey UK

This paper proposes a scheme for multiple un-manned aerial vehicles (UAVs) to track multiple targets in challenging 3-D environments while avoiding obstacle collisions. The scheme relies on Received-Signal-Strength-Indicator (RSSI) measurements to estimate and track target positions and uses a Q-Learning (QL) algorithm to enhance the intelligence of UAVs for autonomous navigation and obstacle avoidance. Considering the limitation of UAVs in their power and computing capacity, a global reward function is used to determine the optimal actions for the joint control of energy consumption, computation time, and tracking accuracy. Extensive simulations demonstrate the effectiveness of the proposed scheme, achieving accurate and efficient target tracking with low energy consumption.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Bio-inspired Swarm Intelligence: a Flocking Project With Group Object Recognition

Bio-inspired Swarm Intelligence: a Flocking Project With Gro...

引用

Consumer Electronics and Computer engineering (ICCECE), 2021 IEEE International Conference on

作者： Song Tianbo Hu Weijun Cai Jiangfeng Liu Weijia Yuan Quan He Kun Ira A. Fulton Schools of Engineering Arizona State University Tempe USA Faculty of Applied Mathematics and Control Processes Saint Petersburg State University Saint Petersburg Russia Department of Chemistry Lomonosov Moscow State University Moscow Russia Allen School of Engineering & Computing Trine University Angola USA Faculty of Engineering The Open University of China Hangzhou China College of Computing Illinois Institute of Technology Chicago USA

With the help of the adaptive gain control algorithm and the flocking SWARM algorithm with adversarial agents, the project described in this article makes the entire system communication-free while maintaining the same functionality and performance. It outlines the entirety of this biologically inspired swarm intelligence's thought process, from inception to conclusion (Flocking algorithm). The entire project is dependable and practically impenetrable under a variety of conditions. The neighbor observation for group conjecture is then strengthened by the use of the adaptive gain control technique and a partial Kalman filter. After the system has been stabilized, use the built-in camera of the SWARM drone to introduce the Object Recognition method in computer vision. As a result, the system enters the real world from the simulation level.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Spatial Temporal Transpose Convolutional Networks for Action Recognition

Spatial Temporal Transpose Convolutional Networks for Action...

引用

International Conference on Computer and Communications (ICCC)

作者： Ying Bai Dongsheng Yang Hongliang Wang Changyi Deng Wei Wang Shenyang Institute of Computing Technology Chinese Academy of Sciences University of Chinese Academy of Sciences Liaoning Province Human-Computer Interaction System Engineering Research Center Based on DigitTwin Shenyang China Shenyang Institute of Computing Technology Chinese Academy of Sciences University of Chinese Academy of Sciences Shenyang China Shenyang Institute of Computing Technology Chinese Academy of Sciences Liaoning Province Human-Computer Interaction System Engineering Research Center Based on DigitTwin Shenyang China Institute of Software China Industrial Control Systems Cyber Emergency Response Team Beijing China

ISBN: (数字)9798331507077

ISBN: (纸本)9798331507084

In recent years, the use of skeleton data for human action recognition has received increasing attention in fields such as intelligent monitoring, human-object interaction, and video retrieval. Among them, Modeling Human Skeleton as Spatial-Temporal Graph Convolutional Network (ST-GCN) has made significant progress. However, methods based on Graph Convolution (GCN) generally face two issues: on one hand, the receptive fields of the spatial and temporal graphs are not flexible enough, limiting their feature representation capabilities. On the other hand, the model is overly parameterized, leading to high computational costs. In this paper, we first propose a spatial-temporal transpose convolution module to overcome the inflexibility of the spatial-temporal graph receptive field. Secondly, we introduce 1 *1 convolutions and depth-wise separable convolution layers into the network model, effectively reducing the computational complexity of the model. On the two subsets of the large-scale skeleton action recognition dataset NTU RGB-D, experimental results show that the proposed method achieved Top1 accuracy metrics of 84.6% and 92.5% respectively, while the model size was reduced by 13.48 times compared to the classic ST-GCN method, and the computational complexity was reduced by approximately 17.18 times.

关键词： Hands Accuracy Convolution Computational modeling Benchmark testing Approximation algorithms Skeleton Partitioning algorithms Computational complexity Monitoring

来源：评论

学校读者我要写书评

暂无评论

Seamless Graph Task Scheduling over Dynamic Vehicular Clouds: A Hybrid Methodology for Integrating Pilot and Instantaneous Decisions

arXiv

引用

arXiv 2025年

作者： Guo, Bingshuo Liwang, Minghui Xia, Xiaoyu Li, Li Jiao, Zhenzhen Hosseinalipour, Seyyedali Wang, Xianbin School of Informatics Xiamen University Fujian China School of Computing Technologies RMIT University MelbourneVIC Australia Department of Control Science and Engineering Tongji University Shanghai China iF-Labs Beijing Teleinfo Technology Co. Ltd. CAICT China Department of Electrical Engineering University at Buffalo-SUNY NY United States Department of Electrical and Computer Engineering Western University ON Canada

Vehicular clouds (VCs) play a crucial role in the Internet-of-Vehicles (IoV) ecosystem by securing essential computing resources for a wide range of tasks. This paPertackles the intricacies of resource provisioning in dynamic VCs for computation-intensive tasks, represented by undirected graphs for parallel processing over multiple vehicles. We model the dynamics of VCs by considering multiple factors, including varying communication quality among vehicles, fluctuating computing capabilities of vehicles, uncertain contact duration among vehicles, and dynamic data exchange costs between vehicles. Our primary goal is to obtain feasible assignments between task components and nearby vehicles, called templates, in a timely manner with minimized task completion time and data exchange overhead. To achieve this, we propose a hybrid graph task scheduling (P-HTS) methodology that combines offline and online decision-making modes. For the offline mode, we introduce an approach called risk-aware pilot isomorphic subgraph searching (RA-PilotISS), which predicts feasible solutions for task scheduling in advance based on historical information. Then, for the online mode, we propose time-efficient instantaneous isomorphic subgraph searching (TE-InstaISS), serving as a backup approach for quickly identifying new optimal scheduling template when the one identified by RA-PilotISS becomes invalid due to changing conditions. Through comprehensive experiments, we demonstrate the superiority of our proposed hybrid mechanism compared to state-of-the-art methods in terms of various evaluative metrics, e.g., time efficiency such as the delay caused by seeking for possible templates and task completion time, as well as cost function, upon considering different VC scales and graph task topologies. Copyright © 2025, The Authors. All rights reserved.

关键词： Cost functions

来源：评论

学校读者我要写书评

暂无评论

Diff9D: Diffusion-Based Domain-Generalized Category-Level 9-DoF Object Pose Estimation

引用

IEEE Transactions on Pattern Analysis and Machine Intelligence 2025年第7期47卷 5520-5537页

作者： Jian Liu Wei Sun Hui Yang Pengchao Deng Chongpei Liu Nicu Sebe Hossein Rahmani Ajmal Mian National Engineering Research Center for Robot Visual Perception and Control Technology College of Electrical and Information Engineering School of Robotics and the State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body Hunan University Changsha China Institute of Artificial Intelligence and Robotics Xi’an Jiaotong University Xi’an China Department of Information Engineering and Computer Science University of Trento Trento Italy School of Computing and Communications Lancaster University Lancaster U.K. Department of Computer Science The University of Western Australia Crawley WA Australia

Nine-degrees-of-freedom (9-DoF) object pose and size estimation is crucial for enabling augmented reality and robotic manipulation. Category-level methods have received extensive research attention due to their potential for generalization to intra-class unknown objects. However, these methods require manual collection and labeling of large-scale real-world training data. To address this problem, we introduce a diffusion-based paradigm for domain-generalized category-level 9-DoF object pose estimation. Our motivation is to leverage the latent generalization ability of the diffusion model to address the domain generalization challenge in object pose estimation. This entails training the model exclusively on rendered synthetic data to achieve generalization to real-world scenes. We propose an effective diffusion model to redefine 9-DoF object pose estimation from a generative perspective. Our model does not require any 3D shape priors during training or inference. By employing the Denoising Diffusion Implicit Model, we demonstrate that the reverse diffusion process can be executed in as few as 3 steps, achieving near real-time performance. Finally, we design a robotic grasping system comprising both hardware and software components. Through comprehensive experiments on two benchmark datasets and the real-world robotic system, we show that our method achieves state-of-the-art domain generalization performance.

关键词： Shape Solid modeling Three-dimensional displays Pose estimation Training Robots Grasping Diffusion models Real-time systems Point cloud compression

来源：评论

学校读者我要写书评

暂无评论

An Optimal Pricing Strategy for Ethereum Gasprice Based on Non-Preemptive Priority Queuing Game

SSRN

引用

SSRN 2023年

作者： Ma, Yu Wang, Xingya Gu, Xin Zhao, Zhihong College of Computer and Information Engineering College of Artificial Intelligence Nanjing Tech University Jiangsu Nanjing211816 China College of Command and Control Engineering Army Engineering University of PLA Jiangsu Nanjing210001 China College of Computing Science and Artificial Intelligence Suzhou City University Jiangsu Suzhou215104 China

In Ethereum, a game-theoretic relationship exists among transactions competing to be included in blocks. Miners prioritize transactions in the memory pool by their gasPrice, and select higher-ranked transactions for confirmation in blocks firstly. Consequently, from the perspective of transactions, this process of selection and packaging can be likened to a queuing system. To assist users in saving transaction fees and reducing waiting time, this paper utilizes a queuing game model to describe the transaction confirmation competition process and proposes an optimal pricing strategy for Ethereum transaction gasPrice that satisfies the Nash equilibrium. Specifically, the paper first models the queuing competition for transaction confirmation as a non-preemptive queuing model with priority. Then, it analyzes the impact of gasPrice on waiting time, deducing a functional relationship between waiting time and gasPrice, thereby deriving the users’ Nash equilibrium payment strategy. Experimental results demonstrate that compared to two state of the art machine learning-based pricing strategies and one based on Poisson regression, applying the pricing strategy proposed in this paper increases the Packing Rates of transactions being included in the next block by 56.21%, 50.95%, and 39.15%, respectively, and achieves cost savings of 21 times, 297 times, and 3 times for users. The results indicate that the gasPrice optimal pricing strategy proposed in this paper helps users effectively balance waiting time and cost. © 2023, The Authors. All rights reserved.

关键词： Game theory

来源：评论

学校读者我要写书评

暂无评论

Model of decentralized cross-chain energy trading for power systems

引用

Global Energy Interconnection 2021年第3期4卷 324-334页

作者： Shaomin Zhang Cong Hou School of Control and Computer Engineering North China Electric Power UniversityEngineering Research Center of Intelligent Computing for Complex Energy SystemsMinistry of EducationHebei Baoding 071003P.R.China

With the development of the energy Internet, more distributed generators are connected to the power grid, resulting in numerous heterogeneous energy networks. However, different energy networks cannot perform efficient energy trading in the centralized management mode, this deeply affecting the complementary ability of heterogeneous energy, resulting in the islanded energy phenomenon. In this model, the same energy on the chain is traded within the chain, and the heterogeneous energy on different chains is traded across chains. To trade energy between heterogeneous energy networks more efficiently, the blockchain-based cross-chain model is proposed based on the existing infrastructure. Heterogeneous energy nodes are assigned to different energy sub-chains and cross-chain energy transactions are performed through a relay-chain, which utilizes the improved Boneh–Lynn–Shacham signature scheme consensus algorithm based on the proof-of-stake and practical Byzantine fault tolerance. The experimental simulations on energy trading efficiency, throughput, and security, show its superiority over existing systems. Further, the simulation results provide a reference for the application of cross-chain technology in energy interconnection.

关键词： Energy internet Blockchain Cross-chain transaction Distributed generation Consensus algorithm

来源：评论

学校读者我要写书评

暂无评论

INTELLIGENT TRACKING AND NAVIGATING MOVING OBJECTS IN A SMART ENVIRONMENT USING IOT NETWORK 1

INTELLIGENT TRACKING AND NAVIGATING MOVING OBJECTS IN A SMAR...

引用

1st International Conference on Technologies for Smart Green Connected Society 2021, ICTSGS 2021

作者： Parvathi, R. Kumar, T.Ch. Anil Nandhakumar, A. Arshad, Haqqani Singh, Ajith B. Department of Information Technology School of Computing Science Hindustan Institute of Technology and Science Padur Chennai India Department of Mechanical Engineering Vignan's Foundation for Science Technology and Research Vadlamudi Dt. Andhra Pradesh Guntur India Department of Electronics and Communication Engineering Dhaanish Ahmed Institute of Technology Chennai India Department of Mechanical Engineering Yanbu Industrial Collge Yanbu Western Region Saudi Arabia Department- Instrumentation and Control Engineering Sri Krishna College of Technology Coimbatore India

ISBN: (纸本)9781607685395

Internet of things (IoT) constitutes the combinations of sensors, controllers, actuators, connectivity. Tracking and navigating moving objects are crucial tasks for security as well as a secure and smart environment. By using the sensory data, location of objects can be identified then that sensory data will be transferred to controllers for the further processes. One pivot role of the system is to identify the trajectory and location of the object using the speed, velocity, acceleration, maps etc and this can be done by machine learning programs. Analyzing service required, prediction of the trajectory and real-world knowledge of maps or locations, To avoid the loop-carry dependency, service request parameters are dynamically changed to locate relevant virtual objects. In order to track moving items, this study provides an architecture based on IoT self-aware environment technology that facilitates the tracking and allocation of relevant virtual objects. © The Electrochemical Society

关键词： IoT loop-carry dependency machine learning object identification self-aware environment sensory data

来源：评论

学校读者我要写书评

暂无评论

Analysis of fast prototyping of microcontroller-based ML software for acoustic signal classification

Analysis of fast prototyping of microcontroller-based ML sof...

引用

Signal Processing Algorithms, Architectures, Arrangements and Applications (SPA)

作者： Filip Marciniak Wojciech Marciniak Tomasz Marciniak Faculty of Computing and Telecommunications Poznan University of Technology Poznań Poland DECYBEL Student Science Club Poznan University of Technology Poznań Poland Faculty of Automatic Control Robotics and Electrical Engineering Institute of Automatic Control and Robotics Division of Electronic Systems and Signal Processing Poznan University of Technology Poznań Poland

The paper analyzes the preparation of software for acoustic signal classification with machine learning techniques for microcontrollers. The design process was tested for three types of devices: Nordic Thingy:53, *** and Arduino Nano 33 BLE Sense Lite. The classifier training process was carried out using the Edge Impulse platform. Experimental studies were carried out for the process of classifying sound signals generated by the vacuum cleaner motor. The results of the training and the model test were presented for different configurations.

关键词：

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：