检索结果-内蒙古大学图书馆

14th International Symposium on Chinese Spoken Language Processing, ISCSLP 2024

作者： Liu, Shuanghong Song, Zhida Fang, Zhihua He, Liang School of Computer Science and Technology Xinjiang University Urumqi China School of Intelligence Science and Technology Xinjiang University Urumqi China Xinjiang Key Laboratory of Signal Detection and Processing Urumqi China Department of Electronic Engineering Beijing National Research Center for Information Science and Technology Tsinghua University Beijing China

ISBN: (纸本)9798331516826

Due to its superior performance and fewer parameters, CAM++ has become the state-of-the-art model for speaker verification tasks. This model uses 2D convolutional blocks to extract front-end features, which are then fed into a densely connected time-delay neural network backbone to extract deep features. However, the simple stacking of 2D convolutions may lead to the generation of a significant amount of redundant features, which is detrimental to efficient feature extraction. Furthermore, although CAM++ already has a relatively small number of parameters, there is still room for further optimization. To address these issues, this paper first employs depthwise separable convolutions to replace the dilated convolutions in the backend network of CAM++, making the model more lightweight. Next, we introduce spatial and channel reconstruction convolution (SCConv) in the ResBlock module of CAM++ to reduce redundant features and optimize the feature extraction process. Finally, after SCConv, we apply squeeze and excitation attention mechanism to model the interdependencies between channels and recalibrate each channel, further enhancing the model’s representational capacity. We name the resulting model LE-CAM++. Our proposed model achieves an EER of 0.686 and a minDCF of 0.084 on the VoxCeleb1-O dataset. Compared to the baseline model CAM++, the EER is reduced by 11%, and the minDCF is reduced by 28%. Additionally, the model parameters are reduced by 8%. ©2024 IEEE.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

UDBNet: Detecting Unsafe Driving Behaviors Relevant to Urban Traffic Safety 30

UDBNet: Detecting Unsafe Driving Behaviors Relevant to Urban...

引用

30th International Conference on Geoinformatics, Geoinformatics 2023

作者： Cheng, Qimin Yang, Yunfei Li, Huanying Ling, Jiajun Shao, Zhenfeng Huazhong University of Science and Technology School of Electronic Information and Communications Wuhan China Tsinghua University Department of Computer Science and Technology Beijing China Wuhan University State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing Wuhan China

ISBN: (纸本)9798350344240

Recently, the number of traffic accident deaths caused by unsafe driving has been increasing dramatically. At present, unsafe driving behavior detection in urban traffic remains a challenging task due to complex background, low spatial resolution, and poor illumination conditions etc., which limit generalization ability and low detection accuracy. In this paper, a novel detection network, called UDBNet, is proposed for unsafe driving behavior detection in complex traffic monitoring scenes. UDBNet introduces a deformable global convolution block to overcome the shortcomings of the existing Convolutional Neural Network (CNN) in capturing global features and adapting to deformed targets. In addition, UDBNet adopts a new Feature Pyramid Network (FPN), termed MBFPN, to enhance the ability of feature fusion. Finally, the effectiveness, robustness, and generalization of UDBNet are proved through extensive experiments. © 2023 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Amplification of high-order azimuthal mode based on a ring-core Yb-doped fiber

引用

Optoelectronics Letters 2022年第4期18卷 222-226页

作者： OU Nanxian LI Wei QIU Runzhou ZHANG Bin GAO Shecheng LIU Weiping Department of Electronic Engineering College of Information Science and TechnologyJinan UniversityGuangzhou 510632China Fiberhome Telecommunication Technologies Co. Ltd.Wuhan 430074China Guangdong Provincial Key Laboratory of Optoelectronic Information Processing Chips and Systems and School of Electronics and Information Technology Sun Yat-sen UniversityGuangzhou 510275China

In order to increase the number of amplified azimuthal modes in Yb-doped fiber(YDF),a multiple azimuthal modes amplifier based on a ring-core Yb-doped fiber(RC-YDF)was proposed and demonstrated.A home-made RC-YDF which can support 6 azimuthal mode groups was employed to amplify the signal mode at 1064 nm,using a core pump *** amplification characteristics of 5 high-order azimuthal linear polarization(HA-LP)mode groups(LP_(11),LP_(21),LP_(31),LP_(41),LP_(51))were studied comprehensively.A more than 8 d B gain is obtained for each signal mode with 5 d Bm input power,and the associated differential modal gain between all modes is less than 1 d *** intensity profiles of all modes are stable and well preserved during the process of amplification.

关键词： mode order modal

来源：评论

学校读者我要写书评

暂无评论

Analysis of Traditional and Deep Learning Architectures in NLP: Towards Optimal Solutions 6

Analysis of Traditional and Deep Learning Architectures in N...

引用

6th International Conference on Recent Trends in Advance Computing, ICRTAC 2023

作者： Subbulakshmi, T. Narayan, Prathiba Lakshmi Suthraye Gokulnath, Shreejith Suganya, R. Subramanian, Girish H. Vellore Institute of Technology School of Computer Science and Engineering Chennai India Pennsylvania State University Information Systems School of Business Administration PA United States

ISBN: (纸本)9798350394702

In the ever-evolving landscape of Natural Language Processing (NLP), the development of novel architectures and optimization techniques has been instrumental in advancing the field. This survey paper presents a comprehensive exploration of traditional and deep learning architectures employed in NLP, while also delving into the optimization strategies that enhance their performance. With a primary focus on surveying existing literature, the aim is to provide a holistic view of the landscape of NLP architectures and their associated optimization techniques. Additionally, a novel architecture is introduced that promises to contribute to the ongoing progress in NLP. This architecture, detailed in our full paper, combines the best practices and innovations from traditional and deep learning models to tackle NLP tasks effectively. Through this work, the aim is to contribute to the collective knowledge in NLP and facilitate future advancements in the field. © 2023 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Zero-DCE++Inspired Object Detection in Less Illuminated Environment Using Improved YOLOv5

引用

Computers, Materials & Continua 2023年第12期77卷 2751-2769页

作者： Ananthakrishnan Balasundaram Anshuman Mohanty Ayesha Shaik Krishnadoss Pradeep Kedalu Poornachary Vijayakumar Muthu Subash Kavitha Centre for Cyber Physical Systems Vellore Institute of Technology(VIT)ChennaiTamil Nadu600127India School of Computer Science and Engineering Vellore Institute of Technology(VIT)ChennaiTamil Nadu600127India School of Information and Data Sciences Nagasaki UniversityNagasaki8528521Japan

Automated object detection has received the most attention over the *** cases ranging from autonomous driving applications to military surveillance systems,require robust detection of objects in different illumination ***-of-the-art object detectors tend to fare well in object detection during daytime ***,their performance is severely hampered in night light conditions due to poor *** address this challenge,the manuscript proposes an improved YOLOv5-based object detection framework for effective detection in unevenly illuminated nighttime ***,the preprocessing strategies involve using the Zero-DCE++approach to enhance lowlight *** is followed by optimizing the existing YOLOv5 architecture by integrating the Convolutional Block Attention Module(CBAM)in the backbone network to boost model learning capability and Depthwise Convolutional module(DWConv)in the neck network for efficient compression of network *** Night Object Detection(NOD)and Exclusively Dark(ExDARK)dataset has been used for this *** proposed framework detects classes like humans,bicycles,and *** demonstrate that the proposed architecture achieved a higher Mean Average Precision(mAP)along with a reduction in model size and total parameters,*** proposed model is lighter by 11.24%in terms of model size and 12.38%in terms of parameters when compared to baseline YOLOv5.

关键词： Object detection deep learning nighttime road scenes YOLOv5 DWConv Zero-DCE++ CBAM

来源：评论

学校读者我要写书评

暂无评论

Active Disturbance Rejection Control Design for the Power on Ultrasonic Transducer 41

Active Disturbance Rejection Control Design for the Power on...

引用

第41届中国控制会议

作者： Fuchun Liu Bin Li Haixing Deng Ming Zeng School of Automation Science and Engineering South China University of Technology Key Laboratory of Autonomous Systems and Networked ControlMinistry of Education Unmanned Aerial Vehicle Systems Engineering Technology Research Center of Guangdong

ISBN: (数字)9789887581536

ISBN: (纸本)9781665482561

The Ultrasonic scalpel has good medical characteristics and is widely used in the medical *** the core component of the ultrasonic scalpel system,the ultrasonic transducer usually needs to control its power for good cutting ***,the mathematical model of the ultrasonic transducer is complex and easily influenced by the load and ***,it is challenging to design a power controller that satisfies all situations with different temperatures and *** this paper,we will analyze the equivalent model of the ultrasonic transducer and simplify the controlled object ***,an ADRC controller,which has the advantage of placing less emphasis on the mathematical model and compensating for the disturbances,will be ***,the performance of the ADRC control system will be tested according to a simulation experiment.

关键词： ultrasonic transducer power control ADRC first-order system

来源：评论

学校读者我要写书评

暂无评论

Flash nanocomplexation (FNC): A new microvolume mixing method for nanomedicine formulation

引用

Chinese Chemical Letters 2024年第10期35卷 39-52页

作者： Keyang Li Yanan Wang Yatao Xu Guohua Shi Sixian Wei Xue Zhang Baomei Zhang Qiang Jia Huanhua Xu Liangmin Yu Jun Wu Zhiyu He Frontiers Science Center for Deep Ocean Multispheres and Earth Systems Key Laboratory of Marine Chemistry Theory and TechnologyMinistry of Education/Sanya Oceanographic InstitutionOcean University of ChinaQingdao/Sanya 266100/572024China College of Chemistry and Chemical Engineering Ocean University of ChinaQingdao 266003China School of Materials Science and Engineering Tsinghua UniversityBeijing 100084China Guangzhou City Polytechnic Guangzhou 510520China National Key Laboratory for the Modernization of Classical and Famous Prescriptions of Chinese Medicine Jiangxi University of Chinese MedicineNanchang330004China Division of Life Science The Hong Kong University of Science and TechnologyHong Kong SAR 999077China

The application of nanotechnologies in formulation has significantly promoted the development of modern medical and pharmacological science, especially for nanoparticle-based drug delivery, bioimaging, and theranostics. The advancement of engineering particle design and fabrication is largely supported by a better understanding of how their apparent characteristics(e.g., size and size distribution, surface morphology, colloidal stability, chemical composition) influence their in vivo biological performance, which raises an urgent need for practical nanoformulation methods. Based on turbulent flow mixing and the self-assembly of molecules in fluids, flash technologies emerged as effective bottom-up fabrication strategies for effective nanoformulation. Among the flash technology family, flash nanocomplexation(FNC) is considered a novel and promising candidate that can promote and optimize formulation processes in a precise spatiotemporal manner, thus obtaining excellent fabrication efficiency, reproducibility and expandability. This review presents an overview of recent advances in fabricating drug-delivery nanoparticles using FNC platforms. Firstly, brief introductions to the basic principles of FNC technology were carried out, followed by descriptions of turbulent microvolume mixers that have significantly promoted the efficiency of FNC-based fabrications. Applications of real formulation cases were then categorized according to the self-assembly-driven interactions(including electrostatic interaction, coordination interaction,hydrogen bonding and hydrophobic interaction) and discussed to reveal the progressiveness of fabricating nanoparticles and discuss how its flexibility will provide advances and replenish the philosophy of nanomedicine formulation. In the end, the commercial potential, current limitations, and prospects of FNC technology for nanoformulation will be summarized and discussed.

关键词： Nanoformulation Nanoparticles Self-assembly Flash nanocomplexation Kinetic control Drug delivery

来源：评论

学校读者我要写书评

暂无评论

Animal Vocal Recognition-Based Breeding Tracking and Disease Warning

引用

Journal of Computers (Taiwan) 2023年第4期34卷 127-143页

作者： Xie, Yinggang Xiao, Yangpeng Peng, Xuewei Liu, Qijia Key Laboratory of Information and Communication Systems Ministry of Information Industry Beijing Information Science and Technology University Beijing100101 China Key Laboratory of the Ministry of Education for Optoelectronic Measurement Technology and Instrument Beijing Information Science and Technology University Beijing100101 China

In this study, we investigate the application of a deep learning framework for the recognition of pig vocalizations. This innovative approach aims to actively monitor and evaluate the diverse states of pigs, with an overarching objective to improve the efficiency of pig farming through prompt identification and resolution of issues. In our comprehensive data collection effort, we carefully gathered a vast assortment of vocal samples from 50 pigs, representative of four distinct states: normal, frightened, coughing, and sneezing. We then meticulously analyzed this vocal data using Mel Frequency Cepstral Coefficients (MFCC). For accurate recognition of pig vocalizations, we devised a fusion model that combines the strengths of Residual Networks (ResNet) and Long Short-Term Memory Networks (LSTM). This model was subsequently tailored, trained, and optimized to meet our specific requirements. Upon rigorous evaluation, we found our model to exhibit exceptional performance in pig vocal recognition tasks, thereby reinforcing the potential of deep learning methodologies in revolutionizing the livestock industry. This research notably underscores the potential of deploying efficient real-time health monitoring systems, offering a promising avenue towards modernizing livestock management practices. © 2023 Computer Society of the Republic of China. All rights reserved.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Shared Mental Models and Boundary Management Activities in New Product Development: A Perspective on Multiteam systems

引用

IEEE Transactions on engineering Management 2024年 71卷 1073-1085页

作者： Lei, Yang Wu, Xiaosong Jiang, James The Department of Information Systems and Management Engineering College of Business Southern University of Science and Technology Shenzhen518055 China The Department of Innovation Management Tsinghua Shenzhen International Graduate School Shenzhen518071 China The Department of Business Administration College of Management National Taiwan University Taipei10617 Taiwan

—Multiteam systems (MTSs) have been increasingly adopted as an organizational form to manage large-scale projects. However, new product development (NPD) in the MTSs structure entails challenges with respect to achieving efficiency due to interteam coordination challenges. Shared mental models (SMMs) within a single team have been found to facilitate intrateam coordination;however, further exploration is needed as to whether SMMs within MTSs can improve interteam coordination among component teams and, if so, what mechanisms underlie this. The present article examines the theoretical underpinning of the impacts of SMMs among component team leaders on MTSs’ NPD efficiency, and further considers the impact of external conditions of requirement uncertainty, the most common external situation during NPD. Using survey data from 110 information technology development MTSs and drawing on boundary management perspective, the results reveal that SMMs among component team leaders positively affect NPD efficiency by managing the MTSs boundary, i.e., through boundary entrainment and boundary buffering activities at the system level. In addition, we find that requirement uncertainty plays a vital role in determining the nature of this relationship. Taken together, our findings highlight the importance of the role of SMMs among component team leaders in facilitating NPD efficiency in MTSs settings. © 2022 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.

关键词： Project management

来源：评论

学校读者我要写书评

暂无评论

Electricity Carbon Quota Trading Scheme based on Certificateless Signature and Blockchain

引用

Computer Modeling in engineering & sciences 2024年第2期138卷 1695-1712页

作者： Xiaodong Yang Runze Diao Tao Liu Haoqi Wen Caifen Wang College of Computer Science and Engineering Northwest Normal UniversityLanzhou730070China Institute of China TelecomWanwei China TelecomWanwei Information Technology Co.Ltd.Lanzhou730030China College of Big Data and Internet Shenzhen Technology UniversityShenzhen518118China Key Laboratory of Cryptography and Data Analytics Northwest Normal UniversityLanzhou730070China

The carbon tradingmarket can promote“carbon peaking”and“carbon neutrality”at low cost,but carbon emission quotas face attacks such as data forgery,tampering,counterfeiting,and replay in the electricity trading *** signatures are a new cryptographic technology that can address traditional cryptography’s general essential certificate requirements and avoid the problem of crucial escrowbased on identity ***,most certificateless signatures still suffer fromvarious security *** present a secure and efficient certificateless signing scheme by examining the security of existing certificateless signature *** ensure the integrity and verifiability of electricity carbon quota trading,we propose an electricity carbon quota trading scheme based on a certificateless signature and *** scheme utilizes certificateless signatures to ensure the validity and nonrepudiation of transactions and adopts blockchain technology to achieve immutability and traceability in electricity carbon quota *** addition,validating electricity carbon quota transactions does not require time-consuming bilinear pairing *** results of the analysis indicate that our scheme meets existential unforgeability under adaptive selective message attacks,offers conditional identity privacy protection,resists replay attacks,and demonstrates high computing and communication performance.

关键词： Electricity carbon trading certificateless signature blockchain forgery attack carbon quota

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：