检索结果-南通市图书馆

QUANTIZATION AND TRAINING OF LOW BIT-WIDTH CONVOLUTIONAL NEURAL NETWORKS FOR OBJECT DETECTION

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Computational Mathematics 2019年第3期37卷 349-359页

作者： penghang yin Shuai Zhang yingyong Qi Jack Xin Department of Mathematics University of CaliforniaLos AngelesCA 90095USA Department of Mathematics University of CaliforniaIrvineCA 92697USA

We presen t LBW-Net,an efficient optimization based method for qua nt ization and training of the low bit-width convolutional neural networks(CNNs).Specifically,we quantize the weights to zero or powers of 2 by minimizing the Euclidean distance between full-precision weights and quantized weights during backpropagation(weight learning).We characterize the combinatorial nature of the low bit-width quantization *** 2-bit(ternary)CNNs,the quantization of N weights can be done by an exact formula in O(N log N)*** the bit-width is 3 and above,we further propose a semi-analytical thresholding scheme with a single free parameter for quantization that is computationally *** free parameter is further determined by network retraining and object detection *** LBW-Net has several desirable advantages over full-precision CNNs,including considerable memory savings,energy efficiency,and faster *** experiments on PASCAL VOC dataset show that compared with its 32-bit floating-point counterpart,the performance of the 6-bit LBW-Net is nearly lossless in the object detection tasks,and can even do better in real world visual scenes,while empirically enjoying more than 4× faster deployment.

关键词： Quantization Low bit width deep neural networks Exact and approximate analytical formulas Network training Object detection

Deep Learning for Real-Time Crime Forecasting and Its Ternarization

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Annals of Mathematics,Series B 2019年第6期40卷 949-966页

作者： Bao WANG penghang yin Andrea Louise BERTOZZI P.Jeffrey BRANTINGHAM Stanley Joel OSHER Jack XIN Department of Mathematics University of CaliforniaLos AngelesWestwoodLos AngelesCA 90095USA Department of Anthropology University of CaliforniaLos AngelesWestwoodLos AngelesCA 90095USA Department of Mathematics University of CaliforniaIrvineIrvineCA 92697USA

Real-time crime forecasting is ***,accurate prediction of when and where the next crime will happen is *** known physical model provides a reasonable approximation to such a complex *** crime data are sparse in both space and time and the signal of interests is *** this work,the authors first present a proper representation of crime *** authors then adapt the spatial temporal residual network on the well represented data to predict the distribution of crime in Los Angeles at the scale of hours in neighborhood-sized *** experiments as well as comparisons with several existing approaches to prediction demonstrate the superiority of the proposed model in terms of ***,the authors present a ternarization technique to address the resource consumption issue for its deployment in real *** work is an extension of our short conference proceeding paper[Wang,B.,Zhang,D.,Zhang,D.H.,et al.,Deep learning for real time Crime forecasting,2017,ar Xiv:1707.03340].

关键词： Crime representation Spatial-temporal deep learning Real-time forecasting Ternarization

Epidemiological Insights into Foodborne Pathogens Through qPCR Exploration of Prevalence—Beijing Municipality,China,January 2022-April 2023

在线全文

学校读者我要写书评

暂无评论

China CDC weekly 2024年第18期6卷 385-389,I0001-I0004页

作者： penghang Zhang Xiaochen Ma Yuzhu Liu Tongyu Wang Shuning Huo Xiaoai Zhang Beijing Key Laboratory of Diagnostic and Traceability Technologies for Food Poisoning Institute for Nutrition and Food HygieneBeijing Center for Disease Prevention and ControlBeijingChina Yanjing Medical College of Capital Medical University BeijingChina

Summary What is already known on this topic?Foodborne diseases present a substantial global health *** diagnostic methods have constraints,but advancements in molecular techniques,like quantitative polymerase chain reaction(qPCR),provide a hopeful *** is added by this report?We examined 1,011 stool samples from individuals suspected of foodborne illnesses.

关键词： solution constraints advancement

Spatial changes and driving factors of lake water quality in Inner Mongolia, China

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Arid Land 2023年第2期15卷 164-179页

作者： REN Xiaohui YU Ruihong LIU Xinyu SUN Heyang GENG Yue QI Zhen ZHANG Zhuangzhuang LI Xiangwei WANG Jun ZHU penghang GUO Zhiwei WANG Lixin XU Jifei Inner Mongolia Key Laboratory of River and Lake Ecology School of Ecology and EnvironmentInner Mongolia UniversityHohhot 010021China Key Laboratory of Mongolian Plateau Ecology and Resource Utilization Ministry of EducationHohhot 010021China

Lakes play important roles in sustaining the ecosystem and economic development in Inner Mongolia Autonomous Region of China,but the spatial patterns and driving mechanisms of water quality in lakes so far remain *** study aimed to identify the spatial changes in water quality and the driving factors of seven lakes(Juyanhai Lake,Ulansuhai Lake,Hongjiannao Lake,Daihai Lake,Chagannaoer Lake,Hulun Lake,and Wulannuoer Lake)across the longitudinal axis(from the west to the east)of Inner ***-scale research was conducted using the comprehensive trophic level index(TLI(Σ)),multivariate statistics,and spatial analysis *** results showed that most lakes in Inner Mongolia were weakly *** dissolved solids and salinity of lake water showed obvious zonation *** and phosphorus were identified as the main pollutants in lakes,with high average concentrations of total nitrogen and total phosphorus being of 4.05 and 0.21 mg/L,*** values of TLI(Σ)ranged from 49.14 to 71.77,indicating varying degrees of lake eutrophication,and phosphorus was the main driver of lake *** lakes of Inner Mongolia could be categorized into lakes to the west of Daihai Lake and lakes to the east of Daihai Lake in terms of salinity and TLI(Σ).The salinity levels of lakes to the west of Daihai Lake exceeded those of lakes to the east of Daihai Lake,whereas the opposite trend was observed for lake trophic *** intensity and mode of anthropogenic activities were the driving factors of the spatial patterns of lake water *** is recommended to control the impact of anthropogenic activities on the water quality of lakes in Inner Mongolia to improve lake ecological *** findings provide a more thorough understanding of the driving mechanism of the spatial patterns of water quality in lakes of Inner Mongolia,which can be used to develop strategies for lake ecosystem protection and water resources management i

关键词： salinity lake eutrophication lake water quality comprehensive trophic level index anthropogenic activities Daihai Lake Inner Mongolia

关于县域法治与县域治理研究的检讨——对既有学术研究文献之梳理与分析

在线全文

学校读者我要写书评

暂无评论

政法论坛 2019年第1期37卷 184-191页

作者：黄鹏航中国政法大学法学院

县域法治与县域治理研究不仅要有理论关注,更要有现实关怀。运用多学科的学术视角,把法治、治理和善治理论融和起来,综合开展县域法治政府、社会治理与自治等理论和实践研究,是实现法治国家的基础。县域法治与县域治理所面临的问题,涉... 详细信息

县域法治与县域治理研究不仅要有理论关注,更要有现实关怀。运用多学科的学术视角,把法治、治理和善治理论融和起来,综合开展县域法治政府、社会治理与自治等理论和实践研究,是实现法治国家的基础。县域法治与县域治理所面临的问题,涉及到县域政府、县域社会以及个人具体生产和生活等各个方面。治理离不开法治,没有法律作为制度保障的治理,不会长治久安。应以法治为前提,尊重县域治理的多样性,限制政府公权力,转变政府治理模式,尊重和保障个人权利,最终促进县域社会有序发展,推进县域法治与县域治理进程。

关键词：县域法治县域治理法治治理

视觉辅助的交叉双旋翼无人直升机自主降落控制系统

在线全文

学校读者我要写书评

暂无评论

光学精密工程 2023年第23期31卷 3517-3529页

作者：席鹏航李京阳印明威宝音贺西天津工业大学机械工程学院天津300387 北京清航紫荆装备科技有限公司北京102101 清华大学航天航空学院北京100084 内蒙古工业大学航空学院呼和浩特010051

为了使交叉双旋翼无人直升机在全球定位系统(Global Positioning System,GPS)信号干扰的情况下实现平稳的自主降落,采用AprilTags视觉基准系统中的TAG36H11图像作为降落地标,并基于该降落地标设计了一种视觉辅助的交叉双旋翼无人直升机... 详细信息

为了使交叉双旋翼无人直升机在全球定位系统(Global Positioning System,GPS)信号干扰的情况下实现平稳的自主降落,采用AprilTags视觉基准系统中的TAG36H11图像作为降落地标,并基于该降落地标设计了一种视觉辅助的交叉双旋翼无人直升机自主降落控制系统。对降落地标进行图像预处理,基于Canny边缘检测、Hough变换直线检测和矩形检测,提取降落地标的特征信息,然后通过相机成像原理和坐标系转换解算出无人直升机的相对位姿信息,针对GPS信号干扰的情况下提出了一种利用Hough变换直线检测解算偏航角的算法,最后设计了一种串级PID控制结构的位姿控制算法,位置控制中设计了一种开方控制器,用于限制无人直升机的极限期望位置和最大加速度。同时,在姿态控制中设计了一种交叉双旋翼无人直升机姿态控制机构,通过6个数字舵机控制倾斜盘的倾角实现交叉双旋翼的周期变距,改变无人直升机的飞行姿态。通过Mission Planner飞控地面站进行仿真验证,将视觉处理系统与飞行控制系统组合,仿真结果表明,视觉辅助的飞行控制系统可以很好地跟踪无人直升机的期望位置。经过实验验证,姿态角度偏移量控制在4°以内,位置偏移量控制在0.05 m以内,可以实现交叉双旋翼无人直升机视觉辅助的自主降落。

关键词：视觉辅助无人直升机交叉双旋翼自主降落位置控制姿态控制

Reinforcement Learning Methods in Robotic Fish: Survey

同方期刊数据库博看期刊评论

在线全文

学校读者我要写书评

暂无评论

Reinforcement Learning Methods in Robotic Fish: Survey

县域法治与治理研究的一个样本——关于昆山经验研究文献的初步总结

第43届中国控制会议

作者： penghang Shuai Haipeng Li Yongkang Luo Liangwei Deng The Laboratory of Cognition and Decision Intelligence for Complex Systems Institute of Automation Chinese Academy of Sciences School of Artificial Intelligence University of Chinese Academy of Sciences State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Chinese Academy of Sciences

Reinforcement learning control has gained increasing attention in the field of biomimetic robotic fish, due to its advantage of universal applicability without prior knowledge of dynamic modeling. In this paper, we review the research work on the application of reinforcement learning in the field of robotic fish. We present and discuss the general model of applying reinforcement learning approaches to robotic fish control, such as the design of environment, state, reward and the selection of reinforcement learning algorithm. Furthermore, we propose to divide the typical tasks of reinforcement learning applied in robotic fish into single control and swarm control, and review the recent advancements separately.

关键词： Reinforcement Learning Control Robotic Fish Single Control Swarm Control

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

法治现代化研究 2019年第2期3卷 144-150页

作者：黄鹏航中国政法大学法学院

昆山是最早践行“县域法治化”的先行者。学术界对昆山的县域法治和县域治理,已经给以不同视角的研究。现有昆山经验的研究,虽有关于治理的相关成果,但以昆山县域治理为研究对象,对昆山经验综合性的跨学科研究,尚待进一步开展。以昆山... 详细信息

昆山是最早践行“县域法治化”的先行者。学术界对昆山的县域法治和县域治理,已经给以不同视角的研究。现有昆山经验的研究,虽有关于治理的相关成果,但以昆山县域治理为研究对象,对昆山经验综合性的跨学科研究,尚待进一步开展。以昆山为样本的县域法治和县域治理研究,应当更加关注昆山经验的全面总结和理论升华,这是一种研究路径的选择。

关键词：昆山经验县域治理县域法治

基于ORB的交叉双旋翼无人直升机视觉跟踪控制系统

在线全文

学校读者我要写书评

暂无评论

南京航空航天大学学报 2023年第6期55卷 966-976页

作者：席鹏航李京阳印明威杨洪伟宝音贺西天津工业大学机械工程学院天津300387 中北大学航空宇航学院太原030051 北京清航紫荆装备科技有限公司北京102101 南京航空航天大学航天学院南京211106 清华大学航空航天学院北京100084 内蒙古工业大学航空学院呼和浩特010051

为了使交叉双旋翼无人直升机在全球定位系统(Global positioning system,GPS)信号不稳定的情况下实现对目标的视觉跟踪,设计了一种交叉双旋翼无人直升机视觉跟踪控制系统。首先,设计了一种视觉模块,对跟踪目标进行图像处理,利用基于方... 详细信息

为了使交叉双旋翼无人直升机在全球定位系统(Global positioning system,GPS)信号不稳定的情况下实现对目标的视觉跟踪,设计了一种交叉双旋翼无人直升机视觉跟踪控制系统。首先,设计了一种视觉模块,对跟踪目标进行图像处理,利用基于方向性快速的特征点检测和旋转BRIEF(Oriented FAST and rotated BRIEF,ORB)描述子特征匹配和矩形检测,获取跟踪目标的中心点像素坐标,通过相机模型和旋转矩阵解算出跟踪目标的中心点实际位置。然后,设计了一种跟踪控制模块,采用串级比例积分微分(Proportional-integral-derivative,PID)加前馈的控制结构,为了限制极限位置和最大加速度,位置控制中采用平方根控制器,姿态控制中设计了一种交叉双旋翼姿态控制机构。之后,通过飞控地面站搭建仿真环境,对添加视觉模块的飞行控制系统进行仿真验证,结果表明,无人直升机可以很好地跟踪到视觉模块输出的期望位置。最后,进行实验验证,在设定最大飞行速度为1 m/s的条件下,姿态角最大为8°左右,位置与期望位置之间的误差最大为0.08 m,可以实现交叉双旋翼无人直升机对目标的视觉跟踪。

关键词：交叉双旋翼无人直升机视觉跟踪特征匹配位置控制姿态控制