检索结果-内蒙古大学图书馆

Performance analysis of deep learning-based object detection algorithms on COCO benchmark: a comparative study

Journal of Engineering and Applied science 2024年第1期71卷 76页

作者： Tian, Jiya Jin, Qiangshan Wang, Yizong Yang, Jie Zhang, shuping sun, Dengxun School of Information Engineering Xinjiang Institute of Technology Aksu Xinjiang843100 China Scientific Research Department Xinjiang Institute of Technology Aksu Xinjiang843100 China

This paper thoroughly explores the role of object detection in smart cities, specifically focusing on advancements in deep learning-based methods. Deep learning models gain popularity for their autonomous feature learning, surpassing traditional approaches. Despite progress, challenges remain, such as achieving high accuracy in urban scenes and meeting real-time requirements. The study aims to contribute by analyzing state-of-the-art deep learning algorithms, identifying accurate models for smart cities, and evaluating real-time performance using the Average Precision at Medium Intersection over Union (IoU) metric. The reported results showcase various algorithms’ performance, with Dynamic Head (DyHead) emerging as the top scorer, excelling in accurately localizing and classifying objects. Its high precision and recall at medium IoU thresholds signify robustness. The paper suggests considering the mean Average Precision (mAP) metric for a comprehensive evaluation across IoU thresholds, if available. Despite this, DyHead stands out as the superior algorithm, particularly at medium IoU thresholds, making it suitable for precise object detection in smart city applications. The performance analysis using Average Precision at Medium IoU is reinforced by the Average Precision at Low IoU (APL), consistently depicting DyHead’s superiority. These findings provide valuable insights for researchers and practitioners, guiding them toward employing DyHead for tasks prioritizing accurate object localization and classification in smart cities. Overall, the paper navigates through the complexities of object detection in urban environments, presenting DyHead as a leading solution with robust performance metrics. © The Author(s) 2024.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Improving learning material repositories using student profiles

引用

soft Computing 2025年第5期29卷 2417-2431页

作者： silva Bravo, Natalie Ferraz Martins, André Ferreira de souza Fonseca Rodrigues, Thales Brito Machado, Marcelo Bernardino, Heder soares Vieira, Alex Borges Barbosa, Hélio José Corrêa de souza, Jairo Francisco LApIC Research Group Federal University of Juiz de Fora Minas Gerais Juiz de Fora36036-900 Brazil Postgraduate Program in Computer Science Federal University of Juiz de Fora Minas Gerais Juiz de Fora36036-900 Brazil Postgraduate Program in Informatics Federal University of the State of Rio de Janeiro Rio de Janeiro22290-255 Brazil Department of Computer Science Federal University of Juiz de Fora Minas Gerais Juiz de Fora36036-900 Brazil

aaaa The adaptive learning community seeks to provide solutions to customize and enhance students’ learning experiences when accessing web-based learning systems. The adaptation usually occurs from the use of learning materials and user information data, which turns the adaptation process highly dependent on the quality of the repositories. Then, the best adaptation a system may offer might still not satisfy the users’ needs. In this work, we propose an approach to assist teachers and stakeholders in understanding repositories’ characteristics and their gaps according to students’ needs. Our approach, first, selects the best sequence of learning materials for each student, which is a well-known problem called Adaptive Curriculum sequencing. Then, based on the selected sequences, we use optimization approaches, such as GRAsP and simulated Annealing, to generate new learning materials possibilities that can improve ACs recommendations. This way, our new approach assists teachers in assembling their learning materials. We have evaluated our approach by comparing it to a traditional approach using a real dataset, and the results are promising. In fact, it is possible to design customized materials using a combination of GRAsP and brute force algorithms on the characteristics of the learning materials. © The Author(s), under exclusive licence to springer-Verlag GmbH Germany, part of springer Nature 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Hybrid Deep learning Approach for ship Navigation in Curved River sections Using PPO and CNN

引用

Informatica (slovenia) 2024年第22期48卷 15-30页

作者： Wan, Jianxia Zhang, sukui Navigation College Jiangsu Maritime Institute Nanjing211170 China Digital Engineering Technology Research Development Center for Maritime Safety and Security Jiangsu Maritime Institute Nanjing211170 China

Curved river sections have complex water flow characteristics and difficulties in maneuvering ships through bends, which pose significant challenges to path planning and ship navigation control. The current path research algorithms still have limitations in dealing with curved and complex waterways. Given this, a convolutional neural network control model based on a hybrid controller and near-end strategy optimization is proposed. This model realizes the path and navigation planning of ships in curved river sections through the hybrid controller. This model utilizes convolutional neural networks to extract channel image features of curved river sections and plans the path through proximal strategy optimization algorithms. In the experiment, high-performance computer processors were used to accelerate the model’s training, and the model was validated in a simulation environment. The results showed that when the research model reached 200 iterations in the simulated curved river section, the average reward value was 0.0323, 19.36% higher than the average reward value of other algorithms. The average instantaneous reward of the research model in path planning was 7.95, which was 3.69 and 1.58 higher than the proximal policy optimization model and the convolutional neural network model based on proximal policy optimization, respectively. The success rate of path planning in complex curved river sections was 82%, significantly higher than the other two algorithms, verifying its effectiveness and superiority in complex path planning tasks. Therefore, this study contributes to improving the safety, efficiency, and economic benefits of ship navigation, and promoting the intelligent and automated growth of the shipping industry © 2024 slovene society Informatika. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Reinforcement learning for optimal tracking of large-scale systems with multitime scales

引用

science China(Information sciences) 2023年第7期66卷 5-29页

作者： Jinna LI Hao NIE Tianyou CHAI Frank L.LEWIs School of Information and Control Engineering Liaoning Petrochemical University State Key Laboratory of Synthetical Automation for Process Industries Northeastern University UTA Research Institute University of Texas at Arlington

This paper aims to solve an optimal tracking control(OTC) problem of large-scale systems with multitime scales and coupled subsystems using singular perturbation(sP) theory and reinforcement learning(RL) techniques. A considerable contribution of this paper is the development of a data-driven sP-based RL method for the OTC of unknown large-scale systems with multitime scales. To achieve this, a multitime scale tracking problem was decomposed into a linear quadratic tracker problem for slow subsystems and a dynamical game problem for fast subsystems using the sP theory. Then, the distributed composite feedback controllers were found using a distributed off-policy integral RL algorithm that uses only measured data from the system in real time. Thus, the operational index can follow its prescribed target value via an approximately optimal approach. Theoretical analysis and proof are presented to demonstrate that the sum of the performances of reduced-order subsystems is approximately equal to the performance of the original large-scale system. Finally, numerical and practical examples are provided to validate the effectiveness of the proposed method.

关键词： reinforcement learning singular perturbation multitime scales data-driven control

来源：评论

学校读者我要写书评

暂无评论

A tactile sensor based on photoelasticity

引用

International Journal of Wireless and Mobile Computing 2024年第3期27卷 281-289页

作者： Tao, Bo Huang, Zhili Zhu, Wenqiong Li, Gongfa Key Laboratory of Metallurgical Equipment and Control Technology Ministry of Education Wuhan University of Science and Technology Hubei Wuhan China Research Centre for Biomimetic Robot and Intelligent Measurement and Control Wuhan University of Science and Technology Hubei Wuhan China

In the new industrial revolution, the market demands integrated talents in new engineering fields. To meet the requirements of the new era for cultivating high-level talents in integrated disciplines, this paper takes the photoelastic tactile sensor as a carrier, developing a task-driven teaching method oriented to student achievement. Combining society’s need with professional knowledge learning and innovation ability cultivation of students, a practical and innovative platform is built to meet the needs of different stages of integrated graduate education for bachelor’s and master’s degrees. Achieving simultaneous enhancement of learning, scientific research and innovation capabilities, it proposes a multi-dimensional assessment mechanism for integrated graduate education, involving long-term assessment of students at different stages and levels. It implements a strict entry and exit policy and comprehensively improves the quality of talent cultivation. The results show very obvious effects, significantly improving the quality of graduates and promoting student development. Copyright © 2024 Inderscience Enterprises Ltd.

关键词： Tactile sensors

来源：评论

学校读者我要写书评

暂无评论

Mastering air combat game with deep reinforcement learning

引用

Defence Technology（防务技术） 2024年第4期34卷 295-312页

作者： Jingyu Zhu Minchi Kuang Wenqing Zhou Heng shi Jihong Zhu Xu Han Department of Precision Instruments Tsinghua UniversityBeijing 100084China Department of Computer Science and Technology Tsinghua UniversityBeijing 100084China Chengdu Aircraft Design&Research Institute Aviation Industry Corporation of ChinaChengdu 610000China

Reinforcement learning has been applied to air combat problems in recent years,and the idea of curriculum learning is often used for reinforcement learning,but traditional curriculum learning suffers from the problem of plasticity loss in neural *** loss is the difficulty of learning new knowledge after the network has *** this end,we propose a motivational curriculum learning distributed proximal policy optimization(MCLDPPO)algorithm,through which trained agents can significantly outperform the predictive game tree and mainstream reinforcement learning *** motivational curriculum learning is designed to help the agent gradually improve its combat ability by observing the agent's unsatisfactory performance and providing appropriate rewards as a ***,a complete tactical maneuver is encapsulated based on the existing air combat knowledge,and through the flexible use of these maneuvers,some tactics beyond human knowledge can be *** addition,we designed an interruption mechanism for the agent to increase the frequency of decisionmaking when the agent faces an *** the number of threats received by the agent changes,the current action is interrupted in order to reacquire observations and make decisions *** the interruption mechanism can significantly improve the performance of the *** simulate actual air combat better,we use digital twin technology to simulate real air battles and propose a parallel battlefield mechanism that can run multiple simulation environments simultaneously,effectively improving data *** experimental results demonstrate that the agent can fully utilize the situational information to make reasonable decisions and provide tactical adaptation in the air combat,verifying the effectiveness of the algorithmic framework proposed in this paper.

关键词： Air combat MCLDPPO Interruption mechanism Digital twin Distributed system

来源：评论

学校读者我要写书评

暂无评论

Application of Machine learning for Aboveground Biomass Modeling in Tropical and Temperate Forests from Airborne Hyperspectral Imagery

引用

FOREsTs 2025年第3期16卷 477-477页

作者： Darko, Patrick Osei Metari, samy Arroyo-Mora, J. Pablo Fagan, Matthew E. Kalacska, Margaret McGill Univ Dept Geog Appl Remote Sensing Lab Montreal PQ H3A 0B9 Canada Natl Res Council Canada Automot & Surface Transportat Ottawa ON K1A 0R6 Canada Natl Res Council Canada Flight Res Lab Ottawa ON K1A 0R6 Canada Univ Maryland Baltimore Cty Dept Geog & Environm Syst Baltimore MD 21250 USA

Accurate operational methods used to measure, verify, and report changes in biomass at large spatial scales are required to support conservation initiatives. In this study, we demonstrate that machine learning can be used to model aboveground biomass (AGB) in both tropical and temperate forest ecosystems when provided with a sufficiently large training dataset. Using wavelet-transformed airborne hyperspectral imagery, we trained a shallow neural network (sNN) to model AGB. An existing global AGB map developed as part of the European space Agency's DUE GlobBiomass project served as the training data for all study sites. At the temperate site, we also trained the model on airborne-LiDAR-derived AGB. In comparison, for all study sites, we also trained a separate deep convolutional neural network (3D-CNN) with the hyperspectral imagery. Our results show that extracting both spatial and spectral features with the 3D-CNN produced the lowest RMsE across all study sites. For example, at the tropical forest site the Tortuguero conservation area, with the 3D-CNN, an RMsE of 21.12 Mg/ha (R-2 of 0.94) was reached in comparison to the sNN model, which had an RMsE of 43.47 Mg/ha (R-2 0.72), accounting for a similar to 50% reduction in prediction uncertainty. The 3D-CNN models developed for the other tropical and temperate sites produced similar results, with a range in RMsE of 13.5 Mg/ha-31.18 Mg/ha. In the future, as sufficiently large field-based datasets become available (e.g., the national forest inventory), a 3D-CNN approach could help to reduce the uncertainty between hyperspectral reflectance and forest biomass estimates across tropical and temperate bioclimatic domains.

关键词： convolutional neural network REDD wavelet scattering continuous wavelet transform deep learning spectra-spatial feature extraction

来源：评论

学校读者我要写书评

暂无评论

Enhancing economic cycle forecasting based on interpretable machine learning and news narrative sentiment

引用

TECHNOLOGICAL FORECAsTING AND sOCIAL CHANGE 2025年 215卷

作者： sun, Weixin Wang, Yong Zhang, Li Chen, Xihui Haviour Hoang, Yen Hai Dongbei Univ Finance & Econ Sch Stat Dalian 116025 Peoples R China Keele Univ Keele Business Sch Keele England Univ Econ HCM City Sch Banking Ho Chi Minh City Vietnam Azerbaijan State Univ Econ Women Researchers Council Baku Azerbaijan

The growing prevalence of uncertainty in global events poses significant challenges to economic cycle forecasting, emphasizing the need for more robust predictive models. This study addresses this gap by developing a novel forecasting framework that integrates multiple uncertainty indices to improve accuracy, stability, and interpretability, particularly during uncertainty shocks. To achieve this, several methodological innovations were implemented. First, news sentiment-based uncertainty indices were incorporated as candidate variables to capture uncertainty dynamics. second, Bayesian least absolute shrinkage and selection operator (Bayesian LAssO) was employed for efficient variable selection, mitigating the curse of dimensionality in small samples. Third, the multi-objective Lichtenberg algorithm (MOLA) was applied to optimize the prediction window size, ensuring model robustness. Additionally, a MOLA-based extreme gradient boosting (MOLA-XGBoost) model was developed to fine-tune hyperparameters across dimensions of prediction accuracy, stability, and directional consistency. Finally, sHapley Additive exPlanations (sHAP) theory was used to enhance model interpretability. This study forecasts China's economic cycle using multiple indicators, demonstrating that the proposed approach consistently delivers accurate and robust predictions even under uncertainty shocks. The findings highlight the crucial role of uncertainty indices in improving economic forecasts, offering new insights and methodologies for predictive modeling in volatile environments.

关键词： Economic cycle Interpretable machine learning News narrative sentiment Uncertainty indices Prediction

来源：评论

学校读者我要写书评

暂无评论

Malware Classification based on Deep learning using Convloutional Neural Network 5

Malware Classification Based on Deep Learning using Convlout...

引用

5th International Conference on IoT based Control Networks and Intelligent systems, ICICNIs 2024

作者： suresh Babu, P. Maheswar Reddy, K. Revanth, Pasupula Reddy, K Maheswar Venkateswara Rao, K sai Kalasaligam Academy of Research and Education Department of Computer Science and Engineering Krishnankoil India

ISBN: (纸本)9798331518097

The increasing sophistication and volume of malware have made it a growing concern in the field of cybersecurity. Traditional detection techniques are becoming less effective as malware continues to evolve, making it harder to identify and mitigate. To address these challenges, deep learning offers a promising solution, particularly due to its advanced capabilities in pattern recognition. This project seeks to harness the power of deep learning to improve malware classification and detection. By applying several deep learning algorithms to a large dataset of malware samples, the research aims to outperform traditional detection methods in both speed and accuracy (99%). Key objectives include developing a highly efficient and adaptable model that can swiftly respond to emerging malware threats, thereby strengthening cyber security defenses. The project will explore the use of modern deep learning techniques such as transfer learning, along with architectures like InceptionNet and MobileNet, to classify malware based on image representations of their signatures. The proposed approach is expected to not only provide better detection rates but also be scalable, enabling real-time detection and adaptation to newly discovered malware strains. Ultimately, the research aims to deliver a more robust, scalable solution for malware detection that can significantly improve cybersecurity systems. © 2024 IEEE.

关键词： Cyber attacks

来源：评论

学校读者我要写书评

暂无评论

Evolutionary optimization of Yagi–Uda antenna design using grey wolf optimizer

引用

Neural Computing and Applications 2025年第6期37卷 4155-4183页

作者： Braik, Malik sheta, Alaa Aljahdali, sultan El-Hefnawi, Fatma Al-Hiary, Heba Elashmawi, Walaa H. Computer Science Department Al-Balqa Applied University Salt Jordan Department of Computer Science Southern Connecticut State University New HavenCT United States Computer Science Department Taif University Taif Saudi Arabia Microwave Department Electronics Research Institute Giza Egypt Department of Computer Information System Al-Balqa Applied University Salt Jordan Department of Computer Science Faculty of Computers & Informatics Suez Canal University Ismailia Egypt Department of Computer Science Faculty of Computer Science Misr International University Cairo Egypt

Meta-heuristic optimization algorithms have become widely used due to their outstanding features, such as gradient-free mechanisms, high flexibility, and great potential for avoiding local optimal solutions. This research explored the grey wolf optimizer (GWO) to find the ideal configuration for a six-element Yagi–Uda antenna. The GWO algorithm adjusted the lengths of the antenna wires and the spacings between them. The goal was to maximize the antenna’s ability to transmit signals (throughput gain). Optimal antenna selection relies on various parameters, including gain, bandwidth, impedance matching, frequency, side-lobe levels, etc. The optimization of a six-element Yagi–Uda antenna presents a challenging engineering design problem due to its multimodal and nonlinear nature. Achieving optimal performance hinges on the intricate interplay between the lengths of the constituent elements and the spacing configurations. To this end, a multiobjective function was adopted to design this antenna. The performance of several meta-heuristic algorithms, including genetic algorithms, biogeography-based optimization, simulated annealing, and grey wolf optimizer, was compared. The GWO-based approach has performed better than its competitors. This optimized antenna design based on GWO reported a gain of 14.21 decibel. Therefore, the GWO-based method optimizes antennas that can be further investigated for other antenna design problems. © The Author(s), under exclusive licence to springer-Verlag London Ltd., part of springer Nature 2024.

关键词： simulated annealing

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：