检索结果-内蒙古大学图书馆

2021 CAA Symposium on Fault Detection, Supervision, and Safety for Technical Processes, SAFEPROCESS 2021

作者： Chen, Runfeng Hu, Yang Hou, Yanze Beijing China Science and Technology on Complex Aviation System Simulation Laboratory 9236 Beijing China

ISBN: (纸本)9781665401159

Prognostics and Health Management (PHM) has been researched on for recent years, and extensive studies have been done on hot topics into this field, such as diagnostics, prognostics, and decision making. However, how to evaluate PHM system architecture is still lacking. This is particularly important when designing avionics system equipped with PHM functions. We propose a general, time-variant architecture model that can simulate PHM system. Performance metrics are calculated to quantitatively evaluate the system architecture. This is useful to find differences of architectures in the earlier design stage in engineering. A case study is presented of the use of architecture modelling to show the effectiveness. And future works are discussed. © 2021 IEEE.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

MilChat: A Large Language Model and Application for Military Equipment

MilChat: A Large Language Model and Application for Military...

引用

Machine Learning and Natural Language Processing (MLNLP), International Conference on

作者： Liu Xue Liu Jie Zhu Peipei Xiang Tao National Key Laboratory of Complex Aviation System Simulation Southwest China Institute of Electronic Technology Chengdu China

ISBN: (数字)9798350354973

ISBN: (纸本)9798350354980

To offer the military personnel a good service of assisting them to get information about enemy's equipment outside Internet, we build a system named MilChat which includes a large language model that has the ability to generate texts and an application which is easy to use. The model of MilChat is fine-tuned by the method of LoRA with military datasets from the basic model, Qwen-7B-Chat which is opensource and has good performance. After fine-tuning, the new model has higher scores than the basic one. Besides, we use the method of Prompt Learning to limit the answers into a standard format. Therefore, it is more accurate and professional than the basic model. This study has also a value of industrial application.

关键词： Accuracy Large language models Machine learning Natural language processing Libraries Internet Personnel Standards Military equipment

来源：评论

学校读者我要写书评

暂无评论

NWD-YOLOv5: A YOLOv5 Model for Small Target Detection Based on NWD Loss 6

NWD-YOLOv5: A YOLOv5 Model for Small Target Detection Based ...

引用

6th International Conference on Robotics, Intelligent Control and Artificial Intelligence, RICAI 2024

作者： Zhou, Liliang Ran, Huaming Xiong, Rongling Tan, Ruijie National Key Laboratory of Complex Aviation System Simulation Chengdu610036 China 10th Research Institute China Electronics Technology Group Corporation Chengdu610036 China School of Mathematics Southwest Jiaotong University Chengdu611756 China

ISBN: (纸本)9798331541699

Target detection technology is vital in modern social management and urban security, with small target detection—accurately identifying tiny objects in large scenes—posing a key challenge for improving surveillance system accuracy and real-time performance. The YOLO series is still the mainstream target detection model, for the YOLOv5 model, as the IoU metric it adopts still uses the geometrical overlap of the detection frame as the basis of the loss calculation, while the experiments have proved that this metric is unable to effectively utilize the small-scale characteristics of small targets. Therefore, improving the loss metric in the existing framework can help the model to pay more attention to the spatial semantic features of small targets in the image in the training phase of the network, respectively, and thus improve the detection effect. In this paper, we propose the NWDYOLOv5 model, i.e., we introduce the NWD loss that measures the difference in the distribution of the detection frames for the existing target detection framework YOLOv5, which ensures that the loss of the small targets is not neglected in the calculation of the forward propagation loss of the model, and we purposefully modify the structure of the feature extraction network in the original framework by adding a feature extraction layer for the detection of the small targets, which enhances the ability of the model to capture the subtle features thus improving the accuracy of small target detection. Through comparative experiments, we evaluated the detection performance of the original YOLOv5 model, the YOLOv5 model using only the NWD loss, the YOLOv5 model with only the additional small-target detection layer, and the full NWD-YOLOv5 model for small-targeted human bodies on the screened COCO2017 public dataset, respectively. The experimental results show that compared to the original YOLOv5 model, the model with only the NWD loss improves the mAP@.5 metrics on small-target detection by 16.5

关键词： Digital elevation model

来源：评论

学校读者我要写书评

暂无评论

The Synthetic Aperture Radar-based Image Augmentation Using Generative Adversarial Network 23

The Synthetic Aperture Radar-based Image Augmentation Using ...

引用

Proceedings of the 2023 8th International Conference on systems, Control and Communications

作者： Kai Zhao Wei Xiong Xiaolan Yu Space Engineering University Science and Technology on Complex Electronic System Simulation Laboratory China

ISBN: (纸本)9798400707810

Synthetic aperture radars (SARs), characterized by their excellent performance, have found wide use in military applications. SAR-based image recognition is an important research area. Neural networks (NNs) have been widely used for image recognition and have demonstrated high recognition accuracy. However, NNs require a significant amount of data for training. In situations with limited training data, it is difficult to obtain satisfactory recognition accuracy using NNs. One approach toward alleviating the problems associated with small training sets is data augmentation. Here, we propose a novel SAR-based image augmentation method: using CycleGAN, aerial images are transformed into SARformat images, thereby allowing expansion of the SAR dataset. To improve the quality of these transformed images, we designed two filters: 1) a pre-filter for selecting images that are suitable for the transformation and 2) a post-filter for selecting suitable images after the transformation; this allows approximating the transformed images as real SAR images. We validated our method on the SSDD dataset. Our experimental results show that compared with the traditional data augmentation method, our method exhibits better recognition performance. With only 300 transformed images, the mean average precision improved from 66% to 76%, using only a small fraction of the SSDD dataset, thus overcoming the problem of small-sized training sets.

关键词： Cycle GAN Image augmentation Synthetic aperture radars

来源：评论

学校读者我要写书评

暂无评论

A Service Mining Scheme Based on Semantic for Internet of Things

引用

Chinese Journal of Electronics 2025年第2期23卷 236-242页

作者： Xinming Li Yan Sun The Science and Technology on Complex Electronic System Simulation Laboratory Academy of Equipment Beijing China The School of Computer Science Beijing University of Posts and Telecommunication Beijing China

As the development of Internet of things (IOT), massive sensors have been deployed as the public infrastructure. With the development of in-depth applications in IOT, service discovery and composition are challenges to the end users. To handle this challenge, we develop a service mining scheme based on semantic for IOT to provide users with interesting composite services. In this scheme, services can be combined and recommended to users actively according to the calculation of service similarity and an updatable semantic database. By the results of service similarity, useless compositions can be filtered out so that energy consumption on service flooding will be reduced. The update strategy of semantic database is also given out, by which the composite services can keep up with time and be more applicative. The benefits of the proposed method are that all operations such as calculating, filtering, and updating are simple enough to be performed in sensor networks.

关键词： Smart buildings Databases Filtering Semantics Public infrastructure Ontologies Sensor systems Internet of Things Floods Intelligent sensors

来源：评论

学校读者我要写书评

暂无评论

A Light Gradient Boosting Machine Based Effectiveness Prediction Method of Space Reconnaissance system

A Light Gradient Boosting Machine Based Effectiveness Predic...

引用

IEEE Advanced Information technology, Electronic and Automation Control Conference (IAEAC)

作者： Jiahuan Liu Ping Jian Science and Technology on Complex Electronic System Simulation Laboratory Space Engineering University Beijing China

To address the problem of effectiveness evaluation of space reconnaissance systems with limited data sets, vast array of indexes and inadequate model interpretability, an interpretable effectiveness prediction method based on Light Gradient Boosting Machine(LightGBM) and the Shapley Additive exPlanations(SHAP) is proposed. The thorough evaluation process is given and the index system of space reconnaissance systems is built. Then LightGBM is used to establish a nonlinear mapping the model between the indexes and the effectiveness. Additionally, the model’s hyperparameters are adjusted using Bayesian optimization to improve accuracy and lessen computational load. Furthermore, the feasibility of the proposed method is validated by modeling and evaluating the performance of the space reconnaissance system under a particular operational scenario. The experimental results show that compared with other mainstream machine learning methods, the proposed method has higher prediction accuracy and can efficiently extract the potential characteristics of the space reconnaissance system.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-View Subspace Clustering Via Simultaneously Cycle Projection and Similarity Learning

Multi-View Subspace Clustering Via Simultaneously Cycle Proj...

引用

International Conference on Wavelet Active Media technology and Information Processing (ICWAMTIP)

作者： Xiang Tao Liu Jie Qiao Wenshen Xiao Siying Ye Mao National Key Laboratory of Complex Aviation System Simulation Southwest China Institute of Electronic Technology Chengdu China University of Electronic Science and Technology of China Chengdu China

ISBN: (数字)9798331519254

ISBN: (纸本)9798331519261

At present, most of the existing methods combine the data from different views mechanically to obtain the latent subspace with complementary information, which not only fails to make full use of the data information from specific perspectives but also brings redundancy. Therefore, we propose a novel multi-view subspace clustering method. It utilizes cycle projection to obtain a subspace construction model. At the same time, by learning the global structure and local manifold structure of this subspace while assigning learnable weights to different views, the consensus information among views and specific information from different views are captured according to the importance of different views. In this way, our method effectively reduces the original noise and redundancy, and further carries out the improvement of complementarity and consistency among views. The proposed method is optimized by using the Augmented Lagrangian Multiplier with Alternating Direction Minimization (ALM- ADM) algorithm. Experiment results on benchmark datasets confirm the effectiveness of our method.

关键词： Manifolds Technological innovation Clustering methods Redundancy Noise Clustering algorithms Information processing Media Minimization Similarity learning

来源：评论

学校读者我要写书评

暂无评论

NUDIF: A Non-Uniform Deployment Framework for Distributed Inference in Heterogeneous Edge Clusters

引用

Future Internet 2025年第4期17卷 168-168页

作者： Li, Peng Qing, Chen Liu, Hao National Key Laboratory of Complex Aviation System Simulation Chengdu610036 China Southwest China Institute of Electronic Technology Chengdu610036 China Beijing100876 China

Distributed inference in resource-constrained heterogeneous edge clusters is fundamentally limited by disparities in device capabilities and load imbalance issues. Existing methods predominantly focus on optimizing single-pipeline allocation schemes for partitioned sub-models. However, such approaches often lead to load imbalance and suboptimal resource utilization under concurrent batch processing scenarios. To address these challenges, we propose a non-uniform deployment inference framework (NUDIF), which achieves high-throughput distributed inference service by adapting to heterogeneous resources and balancing inter-stage processing capabilities. Formulated as a mixed-integer nonlinear programming (MINLP) problem, NUDIF is responsible for planning the number of instances for each sub-model and determining the specific devices for deploying these instances, while considering computational capacity, memory constraints, and communication latency. This optimization minimizes inter-stage processing discrepancies and maximizes resource utilization. Experimental evaluations demonstrate that NUDIF enhances system throughput by an average of 9.95% compared to traditional single-pipeline optimization methods under various scales of cluster device configurations. © 2025 by the authors.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Analysis of node importance of satellite network based on triangular motif

Analysis of node importance of satellite network based on tr...

引用

2023 International Conference on Communication Network and Machine Learning, CNML 2023

作者： Pei, Zhongmin Hu, Boren Luo, Zhangkai Ding, Jie Science and Technology on Complex Electronic System Simulation Laboratory Space Engineering University Beijing101416 China Department of Electronics and Optical Engineering Space Engineering University Beijing101416 China

ISBN: (纸本)9798400716683

Motif provides a local high-order substructure angle for the analysis of network structure evolution, which is of great significance for the study of the interaction between satellite network structure and function. Taking the time scrambling null motif in steady state as a reference, the stability analysis of satellite network structure is carried out from the perspective of the local structure of the motif characteristics;define the temporal degree of satellite network, and propose a satellite network node importance analysis method based on the number of node-constructed triangle motifs, and compare the changes of motifs characteristics under the deliberate attack of temporal degree and based on triangle motifs. simulation experiments are conducted on 46-node GPS constellations containing satellites and ground stations, and it is found that the network with higher initial triangle concentration showed higher stability. In GPS satellite network where the number of node triangle motif is positively correlated with their temporal degree, the attack strategy based on the triangle motif accelerates the attenuation of the triangle subgraph, and the node importance analysis method of the satellite network has higher accuracy. © 2023 ACM.

关键词： Satellites

来源：评论

学校读者我要写书评

暂无评论

Modeling the dynamic alignment of business and information systems via the lens of human-centered architecture evolution

引用

Journal of systems Engineering and Electronics 2021年第4期32卷 881-888页

作者： ZHANG Mengmeng YI Shuanghui CHEN Honghui LUO Aimin LIU Junxian ZHANG Xiaoxue Science and Technology on Information Systems Engineering Laboratory National University of Defense TechnologyChangsha 410072China National Key Laboratory for Complex System Simulation Military Academy of SciencesBeijing 100080China

The complexity of business and information systems(IS)alignment is a growing concern for researchers and practitioners *** extant research on alignment architecture fails to consider the human viewpoint,which makes it difficult to embrace emergent *** paper contributes to the extant literature in the following ***,we combine an enterprise architecture(EA)framework with a human viewpoint to address alignment issues in the architecture design phase;second,we describe a dynamic alignment model by developing a humancentered meta-model that explains first-and second-order changes and their effects on alignment *** paper provides better support for the theoretical research and the practical application of dynamic alignment.

关键词： business and information systems(IS)alignment human viewpoint enterprise architecture(EA) meta-model dynamic alignment

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：