检索结果-内蒙古大学图书馆

International Conference on Parallel and Distributed Systems (ICPADS)

作者： Jian Su Zhenlong Liao Qiankun Mao Zhengguo Sheng Alex X. Liu The School of Software Nanjing University of Information Science and Technology Nanjing China The Hwamei College of Life and Health Sciences Zhejiang Wanli University Ningbo China The School of Computer Nanjing University of Information Science and Technology Nanjing China The Department of Engineering and Design University of Sussex Brighton U.K The Department of Computer Science and Engineering Michigan State University East Lansing USA

With the rapid development of the Internet of Things (IoT) related technologies, the application of digital twins (DT) in industry and healthcare becomes possible. Human activity recognition (HAR) is emerging as a hot research area with great potential in healthcare. Activity recognition systems combined with DT will make it easier to monitor human health conditions to improve the quality of life and happiness with individualized healthcare. In this paper, we design an effective HAR system, called HAR-Net, which uses WiFi time series data collected by sensors to train a deep learning network. Deep learning’s great learning ability is utilized to extract features of various human activities for activity recognition. We built the DT system with Unity, which is combined with the HAR system. In the DT system, real-world physical activities are mapped onto human models. The results of activity prediction can be evaluated in real-time in DT, and warnings can be issued quickly when dangerous activities occur. To make our human activity recognition system more adaptive, we propose a one-shot recognition method based on meta-learning. Specifically, we design a Bi-path basic network that extracts features in the time-domain and frequency-domain, and a meta-learning framework with a classification module and a WiFi metric module. Using datasets from different environments, we conducted various experiments on HAR-Net, and the results proved that our presented method was superior to the baseline network.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Learners’ States Monitoring Method Based on Face Recognition Technology

引用

Procedia computer science 2022年 202卷 172-177页

作者： Xiaoyi Qiao Xiangwei Zheng Xiaofang Sun Shuqin Li Yuang Zhang School of Information Science and Engineering Shandong Normal University Jinan 250014 China Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology Jinan 250014 China

At present, many scholars focus on the learners’ states in online learning, however, a complete real-time monitoring system of learners’ states has not been proposed. In order to better detect the learning process and provide learners with a basis for self-reflection, this paper describes a learners’ states monitoring method based on face recognition technology (LSMFR). A learner monitoring system is developed to collect the learners’ states in real-time through external cameras and record the tired blinking, yawning, head posture, and emotional expression of learners respectively. After a comprehensive analysis, the evaluation score of learners’ states is obtained. In the running process of the system, information feedback is completed by recording the detection content in real-time and then states of learning effectiveness are generated. After learning for a short time, learners check the data record for self-reflection. The simulation test is carried out on the system and the analysis shows that the system achieves the correct analysis of the learners’ states, and generates accurate evaluation scores.

关键词： face recognition real-time monitoring procedure analysis learners’ states

来源：评论

学校读者我要写书评

暂无评论

A Dueling-DDPG Architecture for Mobile Robots Path Planning Based on Laser Range Findings 18th

A Dueling-DDPG Architecture for Mobile Robots Path Planning ...

引用

18th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2021

作者： Zhao, Panpan Zheng, Jinfang Zhou, Qinglin Lyu, Chen Lyu, Lei School of Information Science and Engineering Shandong Normal University Jinan250358 China Shandong Provincial Key Laboratory for Distributed Computer Software Novel Technology Jinan250358 China

ISBN: (纸本)9783030891879

Planning an obstacle-free optimal path presents great challenges for mobile robot applications, the deep deterministic policy gradient (DDPG) algorithm offers an effective solution. However, when the original DDPG is applied to robot path planning, there remains many problems such as inefficient learning and slow convergence that can adversely affect the ability to acquire optimal path. In response to these concerns, we propose an innovative framework named dueling deep deterministic policy gradient (D-DDPG) in this paper. First of all, we integrate the dueling network into the critic network to improve the estimation accuracy of Q-value. Furthermore, we design a novel reward function by combining the cosine distance with the Euclidean distance to improve learning efficiency. Our proposed model is validated by several experiments conducted in the simulation platform Gazebo. Experiments results demonstrate that our proposed model has the better path planning capability even in the unknown environment. © 2021, Springer Nature Switzerland AG.

关键词： Robot applications

来源：评论

学校读者我要写书评

暂无评论

A novel sentiment classification based on 'word-phrase' attention mechanism 24

A novel sentiment classification based on 'word-phrase' atte...

引用

24th IEEE International Conference on Computational science and engineering, CSE 2021

作者： Panga, Guangyao Peng, Guobei Peng, Zizhen He, Jie Yang, Yan Mo, Zhiyi Wuzhou University School of Data Science and Software Engineering Wuzhou543002 China Guangxi University School of Computer Electronics and Information NanNing530004 China WuZhou Vocational College Department of Mechanical and Electrical Engineering Wuzhou543002 China

ISBN: (纸本)9781665416603

With the rapid development of the COVID-19 epidemic, people are prone to panic due to delayed and incomplete information received. In order to quickly identify the sentiments of massive Internet users, it provides a good reference for government agencies to formulate healthy public opinion guidance strategies. This paper proposes a novel sentiment classification based on 'word-phrase' attention mechanism (SC-WPAtt). On the basis of TCN, we propose a shallow feature extraction model based on the word attention mechanism, and a deep extraction model based on the phrase attention mechanism. These models can effectively mine the auxiliary information contained in words, phrases (i.e. combined words) and overall comments, as well as their different contributions, so as to achieve more accurate emotion classification. Experiments show that the performance of the SC-WPAtt method proposed in this paper is better than that of the HN-Att method. © 2021 IEEE.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

Learning from Mistakes: A Comprehensive Review of Knowledge Editing for Large Language Models

Learning from Mistakes: A Comprehensive Review of Knowledge ...

引用

IEEE International Conference on Smart Internet of Things (SmartIoT)

作者： Yanhong Li Chunling Fan Mingqing Huang Chengming Li Artificial Intelligence Research Institute Shenzhen MSU-BIT University Shenzhen China School of Science Harbin Institute of Technology (Shenzhen) Shenzhen China School of Electronic and Communication Engineering Shenzhen Polytechnic University Shenzhen China School of Computer and Software Shenzhen Institute of Information Technology Shenzhen China

ISBN: (数字)9798350366440

ISBN: (纸本)9798350366457

In recent years, there has been a growing recognition that large language models like GPT-4 have the capability to store vast amounts of knowledge and possess extremely powerful abili-ties. However, these models are not immune to making mistakes. The knowledge stored within these models may be erroneous or outdated, leading to incorrect outputs. Consequently, it is imperative to implement methods to edit the knowledge within the models. In this paper, we primarily summarize and organize several key knowledge editing methods that have emerged in recent years. By drawing an analogy to the human process of correcting mistakes, we categorize these methods into three types: rote memorization, learning from examples, and thoroughly understanding. In addition, we briefly discuss the potential positive impact of researches on the knowledge mechanisms of large models on knowledge editing. By summarizing the existing researches on the knowledge of large models, we also briefly explore the role of fully utilizing knowledge mechanisms and resolving knowledge conflicts in enhancing the effectiveness of knowledge editing.

关键词： Reviews Large language models Face recognition Internet of Things Faces Immune system

来源：评论

学校读者我要写书评

暂无评论

Research on Smart Education Big Data Platform Using Neural Network

Research on Smart Education Big Data Platform Using Neural N...

引用

Informatics Education and computer Technology Applications (IECA), International Conference o

作者： Chengfu Wei Zhengxian Chen Jincheng Hu Zhaoqian Yao Wenjie Hou Guodong Gu Business School University of Bristol Bristol UK School of Mechanical Engineering Beijing Institute of Technology Beijing China Yantai Institute Harbin Engineering University Shandong China School of Software and Microelectronics Peking University Beijing China Computer School Beijing Information Science and Technology University Beijing China School of Economics and Finance Shanghai International Studies University Shanghai China

ISBN: (数字)9798350375336

ISBN: (纸本)9798350375343

This paper proposes a novel approach that leverages advances in artificial intelligence and database management to address challenges prevalent in the education industry. The core motivation stems from the need to mitigate issues such as data bias, information overload and resource constraints that are common in educational settings. For this purpose, we have created a comprehensive database specifically for the education sector. Our solutions combine state-of-the-art artificial intelligence technology, including large-scale models and neural networks, with database management systems to efficiently process complex educational information data. It provides a powerful and effective solution to solve the inherent data asymmetry problem in the education industry.

关键词： Industries Accuracy Education Standards organizations Neural networks Decision making Organizations

来源：评论

学校读者我要写书评

暂无评论

Adaptive Packet Scheduling Algorithm for Time-Sensitive Service based on DTP and Reinforcement Learning

Adaptive Packet Scheduling Algorithm for Time-Sensitive Serv...

引用

International Conference on Big Data Computing and Communications (BIGCOM)

作者： Jianing Zhang Jianfeng Guan Tao Feng Xianming Gao School of Computer Science(National Pilot Software Engineering School) Beijing University of Posts and Telecommunications Beijing China Academy of Military Sciences Beijing China

With the rapid development of mobile Internet and Internet of things, a series of time-sensitive services such as video conferencing, cloud games, AR / VR have emerged. In order to meet the above time-sensitive services, IETF has proposed Deterministic Network (DetNet) architecture, which provides an ideal deterministic delay through clock synchronization, zero congestion loss and other mechanisms. Deadline-aware Transport Protocol (DTP) is design to specify the deadline in application layer, and then meets the requirements in transport layer, which can support the time-sensitive services easily. However, the current packet transmission mechanisms are all rule-based and relatively static strategies, which cannot meet the demand of time-sensitive service in dynamic networks. Therefore, this paper proposes an algorithm to dynamically adjust the transmission priority by combining the DTP and reinforcement learning to tackle this issue. More specifically, we design the reward function according to the requirements of DTP, and propose the algorithm of congestion control and packet scheduling in the transport layer. We consider not only the priority but also the service deadline. Comprehensive experiments show that our algorithm performs better in the transmission of time-sensitive services compared to traditional packet scheduling strategies.

关键词： Transport protocols Adaptive systems Scheduling algorithms Heuristic algorithms Reinforcement learning Real-time systems Delays

来源：评论

学校读者我要写书评

暂无评论

Towards Planning Package Delivery Paths by Exploiting the Crowd of Private Cars

Towards Planning Package Delivery Paths by Exploiting the Cr...

引用

IEEE International Conference on High Performance Computing and Communications (HPCC)

作者： Hejing Nie Yiping Wen Tiancai Li Sipeng OuYang Muyang Tian School of Computer Science and Engineering Hunan University of Science and Technology China Hunan Key Laboratory for Service computing and Novel Software Technology Hunan University of Science & Technology

Crowdsourced delivery can provide cost-effective deliveries and it is becoming increasingly important in practice. This paper considers a package delivery problem which exploits relays of private cars to transport packages via their commuter routes. Specifically, we propose a method of planning package delivery paths based on the commuter routes of private cars. Using the commuter routes and time constraints as the reference, it can find optimal delivery paths and assign drivers of private cars to transport packages. The simulation experiment results show the validity and effectiveness of such method.

关键词： Product delivery Planning Automobiles Time factors Relays Task analysis

来源：评论

学校读者我要写书评

暂无评论

Invertible Residual Rescaling Models

arXiv

引用

arXiv 2024年

作者： Li, Jinmin Dai, Tao Zha, Yaohua Luo, Yilu Lu, Longfei Chen, Bin Wang, Zhi Xia, Shu-Tao Zhang, Jingyun Tsinghua Shenzhen International Graduate School Tsinghua University China College of Computer Science and Software Engineering Shenzhen University China Research Center of Artificial Intelligence Peng Cheng Laboratory China Department of Computer Science and Technology Harbin Institute of Technology Shenzhen China WeChat Pay Lab33 Tencent China

Invertible Rescaling Networks (IRNs) and their variants have witnessed remarkable achievements in various image processing tasks like image rescaling. However, we observe that IRNs with deeper networks are difficult to train, thus hindering the representational ability of IRNs. To address this issue, we propose Invertible Residual Rescaling Models (IRRM) for image rescaling by learning a bijection between a high-resolution image and its low-resolution counterpart with a specific distribution. Specifically, we propose IRRM to build a deep network, which contains several Residual Downscaling Modules (RDMs) with long skip connections. Each RDM consists of several Invertible Residual Blocks (IRBs) with short connections. In this way, RDM allows rich low-frequency information to be bypassed by skip connections and forces models to focus on extracting high-frequency information from the image. Extensive experiments show that our IRRM performs significantly better than other state-of-the-art methods with much fewer parameters and complexity. Particularly, our IRRM has respectively PSNR gains of at least 0.3 dB over HCFlow and IRN in the ×4 rescaling while only using 60% parameters and 50% FLOPs. The code will be available at https://***/THUKingmin/IRRM. Copyright © 2024, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

USBDAN: Unsupervised Scale-aware and Boundary-aware Domain Adaptive Network for Gastric Tumor Segmentation

USBDAN: Unsupervised Scale-aware and Boundary-aware Domain A...

引用

Annual International Conference of the IEEE engineering in Medicine and Biology Society (EMBC)

作者： Yongtao Zhang Ning Yuan Bing Liu Aocai Yang Hongwei Yu Kuan Lv Jixin Luan Pianpian Hu Haijun Lei Tianfu Wang Guolin Ma Baiying Lei School of Biomedical Engineering Health Science Centers Shenzhen University Department of Medical Imaging Heping Hospital Affiliated to Changzhi Medical College Changzhi China Department of Radiology China-Japan Friendship Hospital Beijing China School of Computer Science and Software Engineering Shenzhen University Shenzhen China

Accurate segmentation of gastric tumors from computed tomography (CT) images provides useful image information for guiding the diagnosis and treatment of gastric cancer. Researchers typically collect datasets from multiple medical centers to increase sample size and representation, but this raises the issue of data heterogeneity. To this end, we propose a new cross-center 3D tumor segmentation method named unsupervised scale-aware and boundary-aware domain adaptive network (USBDAN), which includes a new 3D neural network that efficiently bridges an Anisotropic neural network and a Transformer (AsTr) for extracting multi-scale features from the CT images with anisotropic resolution, and a scale-aware and boundary-aware domain alignment (SaBaDA) module for adaptively aligning multi-scale features between two domains and enhancing tumor boundary drawing based on location-related information drawn from each sample across all domains. We evaluate the proposed method on an in-house CT image dataset collected from four medical centers. Our results demonstrate that the proposed method outperforms several state-of-the-art methods.

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：