检索结果-内蒙古大学图书馆

DDCM: a decentralized density clustering and its results gathering approach

NEURAL COMPUTING & APPLICATIONS 2023年第35期35卷 24743-24754页

作者： Zou, Lida Shandong Univ Finance & Econ Sch Comp Sci & Technol Jinan 250014 Peoples R China Key Lab Machine Learning & Financial Data Min Univ Jinan 250014 Peoples R China

The use of distributed clustering is an important method of solving large-scale data mining problems. There are still some problems associated with distributed clustering, such as a performance bottleneck on the master node and network congestion caused by global broadcasting. This paper proposes a decentralized clustering method based on density clustering and the content-addressable network technique. It can form a cluster with excellent scalability and load balancing capabilities based on several surrounding nodes. In addition, a method is presented for optimizing the way clustering results are gathered in different application scenarios. Based on our extensive experiments, the proposed approach performs three times better than benchmark algorithms in terms of efficiency and has a stable expanding ratio of about 0.6 for large-scale data sets.

关键词： Distributed clustering Decentralization Results gathering Density clustering Content addressable network

来源：评论

学校读者我要写书评

暂无评论

An Intelligent Improvement Method Of Classroom Cognitive Efficiency Based On Multidimensional Interactive Devices

JOURNAL OF APPLIED SCIENCE AND ENGINEERING

引用

JOURNAL OF APPLIED SCIENCE AND ENGINEERING 2023年第3期26卷 445-454页

作者： Zou, Lida Shandong Univ Finance & Econ Sch Comp Sci & Technol Jinan 250014 Peoples R China Key Lab Machine Learning & Financial Data Min Uni Jinan 250014 Peoples R China

With the development and maturity of intelligent interactive devices, introduction of devices into classroom has become inevitable tendency in order to enhance the teaching effect. At present there exist two difficulties when using devices in engineering. It is hard to get cognitive state of students based on multiple monitoring devices. Moreover, how to interact with students using devices intelligently, personally and automatically is challenging in order to improve student's cognition. To solve the above problems, in the paper we design an intelligent management solution about multidimensional interactive devices and propose an active learning algorithm. We aim to improve the judgment and intervention of students' cognition. The extensive experiments demonstrate that the proposed method performs 21% and 43% better than benchmark algorithm, respectively in prediction accuracy and intervention effect for students' cognition.

关键词： Semi-Supervised learning Deep Network Cluster Multidimensional Interactive Classroom Cognitive Efficiency

来源：评论

学校读者我要写书评

暂无评论

Rapid Discovery Approach of Abnormal Stocks Based on Temporal Convolutional Autoencoder

引用

AUTOMATIC CONTROL AND COMPUTER SCIENCES 2022年第3期56卷 209-220页

作者： Zou, Lida Shandong Univ Finance & Econ Sch Comp Sci & Technol Jinan 250014 Shandong Peoples R China Key Lab Machine Learning & Financial Data Min Uni Jinan 250014 Shandong Peoples R China

Temporal convolutional autoencoder has an important value of application in time-series analysis. In the paper we aim to use temporal convolutional autoencoder to help find out abnormal stocks quickly in the scenario of financial trading market. However, there are still two critical problems to solve during the application of temporal convolutional autoencoder. First, trading data of each stock are multidimensional time-series data, while classical temporal convolutional autoencoder only applies to one-dimension data. Second, stock trading data in a market are huge and their analysis consumes a long time, which contradicts the demand of quick decision in stock trading. To solve the above problems, we improve temporal convolutional autoencoder based on multidimensional sampling, convolution kernel generated by prior knowledge, temporal feature reuse, parallel training on clouds. All the techniques help temporal convolutional autoencoder find abnormal stocks quickly and well. Extended experiments demonstrate that our proposed temporal convolutional autoencoder could raise F1 score to more than seventy percent. The largest time efficiency of finding abnormal stocks can be increased by ninety percent as well.

关键词： temporal convolutional autoencoder anomaly detection a priori knowledge time-series feature reusing parallel training

来源：评论

学校读者我要写书评

暂无评论

Progressive Multi-Resolution Loss for Crowd Counting

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2024年第5期34卷 3232-3244页

作者： Yan, Ziheng Qi, Yuankai Li, Guorong Liu, Xinyan Zhang, Weigang Yang, ming-Hsuan Huang, Qingming Univ Chinese Acad Sci Sch Comp Sci & Technol Key Lab Big Data Min & Knowledge Management Beijing 101408 Peoples R China Univ Adelaide Australian Inst Machine Learning Adelaide SA 5005 Australia Harbin Inst Technol Weihai Weihai 264209 Peoples R China Univ Calif Merced EECS Merced CA 95343 USA

Crowd counting is usually handled in a density map regression fashion, which is supervised via an L2 loss between the predicted density map and ground truth. To effectively regulate models, various improved L2 loss functions have been developed to find a better correspondence between predicted density and annotation positions. In this paper, we propose to predict the density map at one resolution but measure its quality via a derived log-formed loss at multiple resolutions. Unlike existing methods that assume density maps at different resolutions are independent, our loss is obtained by modeling the likelihood function inspired by the relationship of density maps across multi-resolutions. We find that the traditional single-resolution L2 loss is a particular case of our derived log-likelihood. We mathematically prove it is superior to a single-resolution L2 loss. Without bells and whistles, the proposed loss substantially improves several baselines and performs favorably compared to state-of-the-art methods on five crowd counting datasets: NWPU-Crowd, ShanghaiTech A & B, UCF-QNRF, and JHU-Crowd++. The source code and trained models are released at https://***/streamer-AP/PML_***.

关键词： Annotations Kernel Loss measurement Bayes methods Costs Predictive models Bandwidth Crowd counting crowd analysis multi-resolution cascade loss

来源：评论

学校读者我要写书评

暂无评论

learning Hierarchical Modular Networks for Video Captioning

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND machine INTELLIGENCE 2024年第2期46卷 1049-1064页

作者： Li, Guorong Ye, Hanhua Qi, Yuankai Wang, Shuhui Qing, Laiyun Huang, Qingming Yang, ming-Hsuan Univ Chinese Acad Sci Sch Comp Sci & Technol Key Lab Big Data Min & Knowledge Management Beijing 100049 Peoples R China Univ Adelaide Australian Inst Machine Learning Adelaide SA 5005 Australia Chinese Acad Sci Inst Comp Technol Key Lab Intelligent Informat Proc Beijing 100045 Peoples R China Univ Calif Merced Merced CA 95343 USA Yonsei Univ Seoul 03722 South Korea Google Mountain View CA 94043 USA

Video captioning aims to generate natural language descriptions for a given video clip. Existing methods mainly focus on end-to-end representation learning via word-by-word comparison between predicted captions and ground-truth texts. Although significant progress has been made, such supervised approaches neglect semantic alignment between visual and linguistic entities, which may negatively affect the generated captions. In this work, we propose a hierarchical modular network to bridge video representations and linguistic semantics at four granularities before generating captions: entity, verb, predicate, and sentence. Each level is implemented by one module to embed corresponding semantics into video representations. Additionally, we present a reinforcement learning module based on the scene graph of captions to better measure sentence similarity. Extensive experimental results show that the proposed method performs favorably against the state-of-the-art models on three widely-used benchmark datasets, including microsoft research video description corpus (MSVD), MSR-video to text (MSR-VTT), and video-and-TEXt (VATEX).

关键词： Video captioning hierarchical modular network scene-graph reward reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Weakly Supervised Video Individual Counting

Weakly Supervised Video Individual Counting

引用

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Liu, Xinyan Li, Guorong Qi, Yuankai Yan, Ziheng Han, Zhenjun van den Hengel, Anton Yang, ming-Hsuan Huang, Qingming Univ Chinese Acad Sci Beijing Peoples R China Macquarie Univ Sydney NSW Australia UCAS Key Lab Big Data Min & Knowledge Management Beijing Peoples R China Chinese Acad Sci Key Lab Intel Info Proc Inst Comput Tech Beijing Peoples R China Univ Adelaide Australian Inst Machine Learning Adelaide SA Australia Univ Calif Merced Merced CA USA

ISBN: (纸本)9798350353006

Video Individual Counting (VIC) aims to predict the number of unique individuals in a single video. Existing methods learn representations based on trajectory labels for individuals, which are annotation-expensive. To provide a more realistic reflection of the underlying practical challenge, we introduce a weakly supervised VIC task, wherein trajectory labels are not provided. Instead, two types of labels are provided to indicate traffic entering the field of view (inflow) and leaving the field view (outflow). We also propose the first solution as a baseline that formulates the task as a weakly supervised contrastive learning problem under group-level matching. In doing so, we devise an end-to-end trainable soft contrastive loss to drive the network to distinguish inflow, outflow, and the remaining. To facilitate future study in this direction, we generate annotations from the existing VIC datasets SenseCrowd and CroHD and also build a new dataset, UAVVIC. Extensive results show that our baseline weakly supervised method outperforms supervised methods, and thus, little information is lost in the transition to the more practically relevant weakly supervised task. The code and trained model can be found at CGNet.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-level cache management of quantitative trading platform

引用

APPLIED MATHEMATICS AND NONLINEAR SCIENCES 2021年第2期6卷 249-260页

作者： Lida, Zou Alterazi, Hassan A. Hdeib, Roaya Shandong Univ Finance & Econ Sch Comp Sci & Technol Jinan 250014 Peoples R China Univ Shandong Key Lab Machine Learning & Financial Data Min Jinan 250014 Peoples R China King Abdulaziz Univ Fac Comp & Informat Technol Informat Technol Dept Jeddah Saudi Arabia Appl Sci Univ Al Ekir Bahrain

With the rapid development of quantitative trading business in the field of investment, quantitative trading platform is becoming an important tool for numerous investing users to participate in quantitative trading. In using the platform, return time of backtesting historical data is a key factor that influences user experience. In the aspect of optimising data access time, cache management is a critical link Research work on cache management has achieved many referential results. However, quantitative trading platform has its special demands. (1) data access of users has overlapping characteristics for time-series data. (2) This platform uses a wide variety of caching devices with heterogeneous performance. To address the above problems, a cache management approach adapting quantitative trading platform is proposed. It not only merges the overlapping data in the cache to save space but also places data into multi-level caching devices driven by user experience. Our extensive experiments demonstrate that the proposed approach could improve user experience up to >50% compared with the benchmark algorithms.

关键词： multi-level cache cache placement cached data merging user experience function quantitative trading platform

来源：评论

学校读者我要写书评

暂无评论

Matrix factorization with a sigmoid-like loss control

引用

NEUROCOMPUTING 2024年 576卷

作者： Xu, Yuan -Yuan Xiao, Hui Zhang, Heng-Ru Wu, Wei-Zhi min, Fan Southwest Petr Univ Sch Comp Sci & Software Engn Chengdu 610500 Peoples R China Southwest Petr Univ Lab Machine Learning Chengdu 610500 Peoples R China Zhejiang Ocean Univ Sch Informat Engn Zhoushan 316022 Peoples R China Zhejiang Ocean Univ Key Lab Oceanog Big Data Min & Applicat Zhejiang P Zhoushan 316022 Peoples R China

Matrix factorization is one of the fundamental approaches of recommender systems. With the popular L2 loss, learning models tend to overfit significantly deviated predictions. However, predicting the actual rating of 5 as 1 or 2 makes no essential difference in the application. In this paper, we design a sigmoid-like function to control the loss of each individual prediction, which has two advantages. First, it reduces the loss corresponding to significantly deviated predictions. Therefore, the impact of these predictions, some of which may be caused by outliers, is also reduced. Second, it is independent of two classical over -fitting control techniques using regular terms and validation data, respectively. Hence, it can be combined with them to form a more powerful method. Experiments are undertaken on six benchmark datasets in comparison with different losses. Results show that the proposed loss function has good performance in terms of MAE, RMSE, and NDCG, however not so good in terms of HR and MAP.

关键词： Loss function Matrix factorization Over-fitting Recommender system

来源：评论

学校读者我要写书评

暂无评论

Morphological estimation of primary branch length of individual apple trees during the deciduous period in modern orchard based on PointNet plus

引用

COMPUTERS AND ELECTRONICS IN AGRICULTURE 2024年 220卷

作者： Sun, Xiaoming He, Leilei Jiang, Hanhui Li, Rui Mao, Wulan Zhang, Dong Majeed, Yaqoob Andriyanov, Nikita Soloviev, Vladimir Fu, Longsheng Northwest A&F Univ Coll Mech & Elect Engn Yangling 712100 Shaanxi Peoples R China Minist Agr & Rural Affairs Key Lab Agr Internet Things Yangling 712100 Shaanxi Peoples R China Shaanxi Key Lab Agr Informat Percept & Intelligen Yangling 712100 Shaanxi Peoples R China Northwest A&F Univ Shenzhen Res Inst Shenzhen 518000 Guangdong Peoples R China Xinjiang Acad Agr Sci Inst Agr Mechanizat Urumqi 830000 Peoples R China Northwest A&F Univ Coll Hort Yangling 712100 Shaanxi Peoples R China Univ Agr Faisalabad Fac Agr Engn & Technol Faisalabad 38000 Pakistan Financial Univ Govt Russian Federat Dept Data Anal & Machine Learning Moscow 125167 Russia

Primary branch length is an important morphological trait of individual apple tree phenotypes. This study presents a novel method for estimating the primary branch lengths of individual apple trees during the deciduous period by distinguishing their instances, i.e., merging those belonging to the same primary branch based on part segmentation outputs of PointNet++. Firstly, colored and colorless 3D-datasets were prepared for training PointNet++ models. The model with higher overall accuracy (OA), class average accuracy (CAA), and mean intersection-over-union (mIoU) was employed to segment the point cloud of a tree into primary branches (PB), trunk (TK), and end-points of primary branches (EPB) of individual apple trees. Skeletonization was applied to the outputs of the three parts of individual apple trees. Subsequently, each primary branch instance was distinguished by determining its corresponding path and retaining the longest path of the same primary branch only. Finally, the primary branch length was estimated by calculating the sum of Euclidean distances between adjacent points on the corresponding path. Results indicated that adding color to point clouds did not improve segmentation accuracy of PointNet++ on segmenting PB, TK, and EPB with similar color features. The PointNet++ model that was trained without color achieved an OA, CAA, and mIoU of 0.84, 0.83, and 0.70, respectively. The proportion of estimated and ground-truth values of the number of primary branches was 93.64 %. The mean absolute percentage error of estimating primary branch lengths was 12.00 %. These findings demonstrate that the proposed method is promising for high-throughput phenotyping of apple trees.

关键词： Morphological trait Primary branch instance Point cloud Deep learning Part segmentation

来源：评论

学校读者我要写书评

暂无评论

NDMFCS: An automatic fruit counting system in modern apple orchard using abatement of abnormal fruit detection

引用

COMPUTERS AND ELECTRONICS IN AGRICULTURE 2023年第1期211卷

作者： Wu, Zhenchao Sun, Xiaoming Jiang, Hanhui Mao, Wulan Li, Rui Andriyanov, Nikita Soloviev, Vladimir Fu, Longsheng Northwest A&F Univ Coll Mech & Elect Engn Yangling 712100 Shaanxi Peoples R China Minist Agr & Rural Affairs Key Lab Agr Internet Things Yangling 712100 Shaanxi Peoples R China Shaanxi Key Lab Agr Informat Percept & Intelligent Yangling 712100 Shaanxi Peoples R China Northwest A&F Univ Shenzhen Res Inst Shenzhen 518000 Guangdong Peoples R China Financial Univ Govt Russian Federat Dept Data Anal & Machine Learning Moscow 125167 Russia Xinjiang Acad Agr Sci Inst Agr Mechanizat Urumqi 830000 Peoples R China

Automatic fruit counting is an important task for growers to estimate yield and manage orchards. Although many deep-learning-based fruit detection algorithms have been developed to improve performance of automatic fruit counting systems, abnormal fruit detection has often been caused by these algorithms detecting non-target fruits that have similar growth characteristics to target fruits. For abnormal fruit detection, detected fruits in the back row of the tree were defined as DFBRT, while detected fruits on the ground were defined as DFG. Both of them would result in a higher number of fruits counting than the ground truth. This study proposes an automatic fruit counting system called NDMFCS (Normal Detection Matched Fruit Counting System) to solve this problem for improving fruit counting accuracy in modern apple orchard. NDMFCS consists of three sub-systems, i.e. object detection based on You Only Look Once Version 4-tiny (YOLOv4-tiny), abatement of abnormal fruit detection based on threshold, and fruit counting based on trunk tracking and identity document (ID) assignment. YOLOv4tiny was selected to implement detection of fruits and trunks, whose output is confidence and pixel coordinates of detected object. The DFBRT and DFG were abated by thresholds to improve detection performance of fruit. This meant that detected fruits were removed when their distance from camera is further than a distance threshold or the confidence of fruit detection is less than a confidence threshold. Finally, fruit counting was implemented by trunk tracking and ID assignment, where each fruit was assigned a unique tracking ID. Results on 10 sets of original videos indicated that average fruit detection precision was improved from 89.1% to 93.3% after abatement of abnormal fruit detection. Also, Multiple Object Tracking Accuracy and Multiple Object Tracking Precision were improved on average by 4.2% and 3.3%, respectively, while average ID Switch Rate was decreased on average by 1.1%.

关键词： Apple orchard Abatement of abnormal fruit detection Object detection Fruit counting ID assignment

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：