检索结果-内蒙古大学图书馆

A Lightweight ST-YOLO Based Model for Detection of Tea Bud in Unstructured Natural Environments

IAENG International Journal of Applied Mathematics 2024年第3期54卷 342-349页

作者： Wen, Xin Yao, Yi Cai, Ying Zhao, Zixing Chen, Tianjiao Zeng, Ziyu Tang, Zhen Gao, Fangzheng the School of Automation of Nanjing Institute of Technology Nanjing211167 China the School of Automation Nanjing Institute of Technology Nanjing211167 China the School of Computer engineering Nanjing Institute of Technology Nanjing211167 China the School of Automation Nanjing Institute of Technology Nanjing211167 China

To address the difficulty of tea shoot recognition in natural environments, an enhanced YOLOX-Nano model(denoted as ST-YOLO) is introduced in this paper. In the proposed algorithm, the Depthwise Separable Convolution is incorporated into the backbone network to alleviate the feature extraction workload. And the Swin Transformer structure is also integrated into the neck network to augment feature extraction capabilities and to enhance the recognition of tea leaf shoots across diverse environmental conditions. Furthermore, to expedite model convergence, the SIoU loss function is employed in localization loss computation. Experimental results shows the efficacy of the proposed ST-YOLO algorithm, with achieved Average Precision (AP) and F1 scores of 86.2% and 88%, respectively. Comparing ST-YOLO with the baseline YOLOX-Nano in the same setting, it reveals notable improvements of 3.3% in AP and 5% in F1. Additionally, compared with the widely adopted target detection algorithms like YOLO v5-s, YOLO v4, and Faster R-CNN, ST-YOLO exhibits the superiority of a remarkable reduction in model size by half while simultaneously enhancing AP values by 6.0%, 6.1%, and 6.88%, respectively. As a result, the proposed algorithm not only excels in terms of model lightweights but also significantly improves detection accuracy, which offers a promising solution for the deployment of intelligent tea-picking systems on embedded hardware platforms. © (2024), (International Association of Engineers). All Rights Reserved.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

A Software Defect Prediction Method Using a Multivariate Heterogeneous Hybrid Deep Learning Algorithm

引用

computers, Materials & Continua 2025年第2期82卷 3251-3279页

作者： Qi Fei Haojun Hu Guisheng Yin Zhian Sun College of Computer Science and Technology Harbin Engineering UniversityHarbin150001China Information Technology Research Department Jiangsu Automation Research InstituteLianyungang222062China School of Computer Science and Technology Harbin Institute of TechnologyShenzhen518055China

Software defect prediction plays a critical role in software development and quality assurance processes. Effective defect prediction enables testers to accurately prioritize testing efforts and enhance defect detection efficiency. Additionally, this technology provides developers with a means to quickly identify errors, thereby improving software robustness and overall quality. However, current research in software defect prediction often faces challenges, such as relying on a single data source or failing to adequately account for the characteristics of multiple coexisting data sources. This approach may overlook the differences and potential value of various data sources, affecting the accuracy and generalization performance of prediction results. To address this issue, this study proposes a multivariate heterogeneous hybrid deep learning algorithm for defect prediction (DP-MHHDL). Initially, Abstract Syntax Tree (AST), Code Dependency Network (CDN), and code static quality metrics are extracted from source code files and used as inputs to ensure data diversity. Subsequently, for the three types of heterogeneous data, the study employs a graph convolutional network optimization model based on adjacency and spatial topologies, a Convolutional Neural Network-Bidirectional Long Short-Term Memory (CNN-BiLSTM) hybrid neural network model, and a TabNet model to extract data features. These features are then concatenated and processed through a fully connected neural network for defect prediction. Finally, the proposed framework is evaluated using ten promise defect repository projects, and performance is assessed with three metrics: F1, Area under the curve (AUC), and Matthews correlation coefficient (MCC). The experimental results demonstrate that the proposed algorithm outperforms existing methods, offering a novel solution for software defect prediction.

关键词： Software defect prediction multiple heterogeneous data graph convolutional network models based on adjacency and spatial topologies CNN-BiLSTM TabNet

来源：评论

学校读者我要写书评

暂无评论

Overhead-free Noise-tolerant Federated Learning: A New Baseline

引用

Machine Intelligence Research 2024年第3期21卷 526-537页

作者： Shiyi Lin Deming Zhai Feilong Zhang Junjun Jiang Xianming Liu Xiangyang Ji Department of Computer Science and Technology Harbin Institute of TechnologyHarbin150000China Department of Automation Tsinghua UniversityBeijing100084China

Federated learning (FL) is a promising decentralized machine learning approach that enables multiple distributed clients to train a model jointly while keeping their data private. However, in real-world scenarios, the supervised training data stored in local clients inevitably suffer from imperfect annotations, resulting in subjective, inconsistent and biased labels. These noisy labels can harm the collaborative aggregation process of FL by inducing inconsistent decision boundaries. Unfortunately, few attempts have been made towards noise-tolerant federated learning, with most of them relying on the strategy of transmitting overhead messages to assist noisy labels detection and correction, which increases the communication burden as well as privacy risks. In this paper, we propose a simple yet effective method for noise-tolerant FL based on the well-established co-training framework. Our method leverages the inherent discrepancy in the learning ability of the local and global models in FL, which can be regarded as two complementary views. By iteratively exchanging samples with their high confident predictions, the two models “teach each other” to suppress the influence of noisy labels. The proposed scheme enjoys the benefit of overhead cost-free and can serve as a robust and efficient baseline for noise-tolerant federated learning. Experimental results demonstrate that our method outperforms existing approaches, highlighting the superiority of our method.

关键词： Federated learning noise-label learning privacy-preserving machine learning edge intelligence distributed machine learning

来源：评论

学校读者我要写书评

暂无评论

Lightweight Rice Planthopper Identification Method Based on YOLOv5

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer Science 2024年第8期51卷 1079-1085页

作者： Li, Siquan Wang, Yi Shi, Teng Chen, Xi Tang, Zhen Zeng, Ziyu Wen, Xin Shang, Yanling School of Automation Nanjing Institute of Technology Nanjing211167 China School of Mechanical Engineering Nanjing Institute of Technology Nanjing211167 China School of Automation Nanjing Institute of Technology Nanjing211167 China School of Computer engineering Nanjing Institute of Technology Nanjing211167 China School of Automation Nanjing Institute of Technology Nanjing211167 China

The identification and classification of pests in rice field are the prerequisites of early warning systems for pest disasters. Among these pests, the rice planthoppers cause the most serious damage. However, the existing deep learning models for rice planthopper recognition are characterized by large size and numerous parameters, which makes them unsuitable for the deployment on embedded devices with limited computational resources. To address this issue, a lightweight rice planthopper recognition model based on YOLOv5 is proposed in this paper. In the model, a lightweight convolutional network named GhostNet is employed as the backbone to reduce the operational parameters. Additionally, a convolutional attention module (CBAM) is integrated into the backbone network to effectively enhance the transmission of deep information, so as to improve the model’s ability of recognizing rice planthopper images. The original CIoU loss function is replaced by the SIoU loss function to expedite model convergence. Experimental results demonstrate that the modified model achieves the mAP@0.5 as 82.8%, with the parameter count of 3.12×106 and the model size of 7.2MB. Compared to the original model, it is a reduction of 46.7% in size and 43.3% in parameters, with a minor accuracy loss of 0.1%. Clearly, the improved model can achieve lightweight characteristics and robust performance, and thus it provides a theoretical and practical foundation for early warning systems against rice planthopper infestations in rice fields. © (2024), (International Association of Engineers). All Rights Reserved.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Design of Large-Scale Hybrid, Hydrogen and Battery, and Energy Storage Systems for Grid Applications

IEEE Open Access Journal of Power and Energy

引用

IEEE Open Access Journal of Power and Energy 2025年 12卷 341-352页

作者： Dorn, Marvin Lotze, Jonas Këhnapfel, Uwe Weber, André Hagenmeyer, Veit Institute for Automation and Applied Informatics Karlsruhe Institute of Technology Eggenstein-Leopoldshafen76344 Germany Institute for Automation and Applied Informatics TransnetBW GmbH Stuttgart70191 Germany Institute for Applied Computer Science Karlsruhe Institute of Technology Eggenstein-Leopoldshafen76344 Germany

Due to the energy transition, which involves phasing out base load power plants such as coal, there is a need to establish storage systems within the energy system to compensate for fluctuations of renewable energies. Batteries are suitable for day-night cycles and particularly for short-cycle applications. To address the problem of dark-doldrums, when neither wind nor solar energy is available, gas and, in the more distant future, hydrogen power plants are to be used. By combining batteries and hydrogen power plants in a hybrid energy storage system, further advantages and application possibilities arise regarding grid stability and system design. This work illustrates interrelationships between the subsystems, optimizes proportions, and demonstrates logical system sizes, technologies, and their costs. A central part of the work are the self-derived methods for system design and the justification of these. Storage pressure, running times, availability time, annual cycles and design of the subsystems are described. Systems of this scale are difficult to imagine. A program developed as part of this work to implement the methods, visualizes the system, displays the system parameters, and shows the best-case and worst-case capital expenditures. An optimized system design is presented. Different combinations in the system design show the effects on capital expenditures. Starting from 2 to 4 hours of availability time, the hybrid system becomes cheaper than a pure battery system in terms of capital expenditures. © 2020 IEEE.

关键词： Electric power transmission networks

来源：评论

学校读者我要写书评

暂无评论

Autonomous embodied navigation task generation from natural language dialogues

引用

Science China(Information Sciences) 2025年第5期68卷 119-132页

作者： Haifeng XU Yongchang LI Lumeng MA Chunwen LI Yanzhi DONG Xiaohu YUAN Huaping LIU Department of Automation Tsinghua University School of Physics and Electronic Information Yantai University School of Electrical and Electronic Engineering Shanghai Institute of Technology Department of Computer Science and Technology Tsinghua University

Robots are increasingly being deployed in densely populated environments, such as homes, hotels, and office buildings, where they rely on explicit instructions from humans to perform tasks. However, complex tasks often require multiple instructions and prolonged monitoring, which can be time-consuming and demanding for users. Despite this, there is limited research on enabling robots to autonomously generate tasks based on real-life scenarios. Advanced intelligence necessitates robots to autonomously observe and analyze their environment and then generate tasks autonomously to fulfill human requirements without explicit commands. To address this gap, we propose the autonomous generation of navigation tasks using natural language dialogues. Specifically, a robot autonomously generates tasks by analyzing dialogues involving multiple persons in a real office environment to facilitate the completion of item transportation between various *** propose the leveraging of a large language model(LLM) through chain-of-thought prompting to generate a navigation sequence for a robot from dialogues. We also construct a benchmark dataset consisting of 625 multiperson dialogues using the generation capability of LLMs. Evaluation results and real-world experiments in an office building demonstrate the effectiveness of the proposed method.

关键词： proactive robot robot navigation service robot large language model

来源：评论

学校读者我要写书评

暂无评论

Short-Term Power Load Forecasting Method Based on GRU-Transformer Combined Neural Network Model

引用

Journal of Computing and Information technology 2024年第1期32卷 1-14页

作者： Mao, Weiwei Yu, Suping Chen, Wenqing College of Computer and Information Engineering Luoyang Institute of Science and Technology Luoyang China School of Electrical Engineering and Automation Luoyang Institute of Technology Luoyang China

Load Forecast (LF) is an important task in the planning, control and application of public power systems. Accurate Short Term Load Forecast (STLF) is the premise of safe and economical operation of a power system. In the research of short-term power load forecasting, machine learning and deep learning are the most popular methods at present, but there still exists a problem that the single and simple structure of power load forecasting model leads to low accuracy of load forecasting. In order to improve the accuracy of STLF, a Gated Cycle Unit (GRU)-Transformer combined neural network model is proposed. Transformer encoder structure is used as feature extractor to mine the complex mapping relationships between the input features and load. The advantage of self-attention mechanism is used to solve the problem of information loss of long sequences in short-term power load forecasting. At the same time, the multivariate time series model of GRU is used for model training. The experimental results on the power load data set of a certain region in southwest China and Panama City show that the proposed combined model prediction method has higher accuracy than those proposed in other literatures, which further proves its feasibility and superiority. © 2024, University of Zagreb Faculty of Electrical Engineering and Computing. All rights reserved.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

Reinforcement learning-based unknown reference tracking control of HMASs with nonidentical communication delays

引用

Science China(Information Sciences) 2023年第7期66卷 46-57页

作者： Yong XU Zheng-Guang WU Wei-Wei CHE Deyuan MENG School of Automation Beijing Institute of Technology Institute of Cyber-Systems and Control Zhejiang University College of Mathematics and Computer Science Zhejiang Normal University Department of Automation Qingdao University School of Automation Science and Electrical Engineering Beihang University(BUAA)

This paper focuses on the optimal output synchronization control problem of heterogeneous multiagent systems(HMASs) subject to nonidentical communication delays by a reinforcement learning *** with existing studies assuming that the precise model of the leader is globally or distributively accessible to all or some of the followers, the leader's precise dynamical model is entirely inaccessible to all the followers in this paper. A data-based learning algorithm is first proposed to reconstruct the leader's unknown system matrix online. A distributed predictor subject to communication delays is further devised to estimate the leader's state, where interaction delays are allowed to be nonidentical. Then, a learning-based local controller, together with a discounted performance function, is projected to reach the optimal output synchronization. Bellman equations and game algebraic Riccati equations are constructed to learn the optimal solution by developing a model-based reinforcement learning(RL) algorithm online without solving regulator equations, which is followed by a model-free off-policy RL algorithm to relax the requirement of all agents' dynamics faced by the model-based RL algorithm. The optimal tracking control of HMASs subject to unknown leader dynamics and communication delays is shown to be solvable under the proposed RL algorithms. Finally, the effectiveness of theoretical analysis is verified by numerical simulations.

关键词： heterogeneous multiagent systems HMAS reinforcement learning RL optimal output synchronization communication delays

来源：评论

学校读者我要写书评

暂无评论

OCRBench: on the hidden mystery of OCR in large multimodal models

引用

Science China(Information Sciences) 2024年第12期67卷 23-35页

作者： Yuliang LIU Zhang LI Mingxin HUANG Biao YANG Wenwen YU Chunyuan LI Xu-Cheng YIN Cheng-Lin LIU Lianwen JIN Xiang BAI School of Artificial Intelligence and Automation Huazhong University of Science and Technology School of Electronic and Information Engineering South China University of Technology Microsoft Research School of Computer & Communication Engineering University of Science and Technology Beijing Institute of Automation Chinese Academy of Sciences School of Software Engineering Huazhong University of Science and Technology

Large models have recently played a dominant role in natural language processing and multimodal vision-language learning. However, their effectiveness in text-related visual tasks remains relatively unexplored. In this paper, we conducted a comprehensive evaluation of large multimodal models, such as GPT4V and Gemini, in various text-related visual tasks including text recognition, scene text-centric visual question answering(VQA), document-oriented VQA, key information extraction(KIE), and handwritten mathematical expression recognition(HMER). To facilitate the assessment of optical character recognition(OCR) capabilities in large multimodal models, we propose OCRBench, a comprehensive evaluation benchmark. OCRBench contains 29 datasets, making it the most comprehensive OCR evaluation benchmark available. Furthermore, our study reveals both the strengths and weaknesses of these models, particularly in handling multilingual text, handwritten text, non-semantic text, and mathematical expression *** importantly, the baseline results presented in this study could provide a foundational framework for the conception and assessment of innovative strategies targeted at enhancing zero-shot multimodal *** evaluation pipeline and benchmark are available at https://***/Yuliang-Liu/Multimodal OCR.

关键词： large multimodal model OCR text recognition scene text-centric VQA document-oriented VQA key information extraction handwritten mathematical expression recognition

来源：评论

学校读者我要写书评

暂无评论

A trajectory-level robot skill learning method based on fuzzy systems

引用

Science China(Information Sciences) 2025年第5期68卷 38-52页

作者： Chao ZENG Zhen DENG Boyang TI Yongsheng GAO Department of Computer Science University of Liverpool School of Mechanical Engineering and Automation Fuzhou University State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences State Key Laboratory of Robotics and System Harbin Institute of Technology

Learning from demonstration(LfD) allows for the effective transfer of human manipulation skills to a robot by building a model that represents these skills based on a limited number of demonstrated ***,a skilllearning model that can comprehensively satisfy multiple requirements,such as computational complexity,modeling accuracy,trajectory smoothness,and robustness,is still ***,this work aims to provide such a model by employing fuzzy ***,we introduce an LfD model named Takagi-Sugeno-Kang fuzzy system-based movement primitives(TSKFMPs),which exploits the advantages of the fuzzy theory for effective robotic imitation learning of human *** work formulates the TSK fuzzy system and gradient descent(GD) as imitation learning models,leveraging recent advancements in GD-based optimization for fuzzy *** study takes a two-step strategy.(ⅰ) The input-output relationships of the model are established using TSK fuzzy systems based on demonstration *** this way,the skill is encoded by the model parameter in the latent space.(ⅱ) GD is used to optimize the model parameter to increase the modeling accuracy and trajectory *** further explain how learned trajectories are adapted to new task scenarios through local *** conduct multiple tests using an open dataset to validate our method,and the results demonstrate performance comparable with those of other ***,we implement it in a real-world case study.

关键词： robot skill learning movement primitives TSK fuzzy systems gradient descent robotics

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：