检索结果-内蒙古大学图书馆

21st IEEE International Conference on High Performance Computing and Communications, 17th IEEE International Conference on Smart City and 5th IEEE International Conference on Data Science and systems, HPCC/SmartCity/DSS 2019

作者： Zuo, Peixuan Wang, Rui Fu, Xianya Yang, Hailong Liu, Yi Zhang, Lianyi Zhang, Han Qian, Depei BeiHang University Beijing China Secience and Technology on Special System Simulation Laboratory Beijing China Beijing Simulation Center Beijing China

ISBN: (纸本)9781728120584

Convolutional Neural Networks have received considerable attention over the past few years, and they are widely used in various fields. However, the computational complexity and excessive storage space caused by over-parameterized of these networks are also being serious. This will limit further development of deep learning in resource-constrained devices such as the mobile terminal. It has been proposed that pruning is a useful method to optimize networks, whereas there is no contribution focus on utilizing the traits of the network to guide pruning. The primary purpose of this paper is to research the different effects of layers with varying depth and the different number of parameters during the pruning process. By taking advantage of the characteristics of network structure itself, we propose a new pruning strategy with three variants to effectively prune CNNs which can automatically determine the most appropriate sparsity for each conv-layer. Using iterative pruning and retraining, we evaluated our strategy on InceptionV3 and ResNet50. We analyzed our three variants and made a comparison between each other. © 2019 IEEE.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

Scattering Center Modeling Using Adaptive Segmental Compressive Sampling

引用

Journal of Beijing Institute of technology 2017年第4期26卷 484-493页

作者： Qifeng Li Kunyi Guo Jiaxin Wang Xinqing Sheng Tianshu Liu Center for Electromagnetic Simulation School of Information and ElectronicsBeijing Institute of Technology Science and Technology on Special System Simulation Laboratory Beijing Simulation Center

In order to deal with aliasing distortions of Doppler frequencies shown in time-frequency representation（ TFR） with aspect undersampling,an approach using adaptive segmental compressive sampling according to the aspect dependencies of the scattering centers is proposed. The random noise problem induced by compressive sampling is solved by employing a series of signal processing techniques of filtering,image transformation and Hough Transform. Three examples are presented to verify the effectiveness of this approach. The comparisons between the built models and the precise scattered fields computed by a well-validated full-wave numerical method are investigated,and the results showgood agreements between each other.

关键词： scattering center undersampling compressive sensing time-frequency representation parameter estimation

来源：评论

学校读者我要写书评

暂无评论

Research on Credibility Evaluation for Radio Frequency Guided Hardware-In-the-Loop simulation system 36

Research on Credibility Evaluation for Radio Frequency Guide...

引用

第36届中国控制会议

作者： Yuxiao Wang Tao Chao Chaolei Wang Songyan Wang Harbin Institute of Technology Science and Technology on Special System Simulation Laboratory Beijing Simulation Center

ISBN: (纸本)9781538629185

To ensure the effectiveness of the RF（Radio Frequency） guided HILS（Hardware-In-the-Loop simulation） system,the credibility evaluation was *** present,most of the methods were applied in the case of the simulation system input being consistent with the real system input,which is difficult to achieve in some subsystem of the RF guided HILS *** this paper,the relative complete system reliability evaluation index system was *** were classified,for different categories of subsystem,corresponding reliability assessment methods were *** particular,reliability evaluation methods based on performance level were put forward,when certain subsystems cannot be evaluated by consistency analysis methods.A simple example was given to show the evaluation process.

关键词： Radio Frequency Guided Hardware-In-the-Loop simulation Credibility Evaluation Index system Performance Index

来源：评论

学校读者我要写书评

暂无评论

A Multi-Targets simulation Method Based on Single Channel

A Multi-Targets Simulation Method Based on Single Channel

引用

2023 International Conference on Electronic Information Engineering and Data Processing, EIEDP 2023

作者： Jing, Boda Jin, Congjun Song, Youbin Beijing Simulation Center Science and Technology on Special System Simulation Laboratory Beijing100854 China

ISBN: (纸本)9781510666160

The number of channels for radio frequency simulation system is very limited, and the traditional simulation method of one channel simulating one target is difficult to meet the demand of complex scene simulation. This paper proposes a simulation method based on the triadic vector control algorithm of "both amplitude and phase control" to simulate the angular position of multiple radar targets by one channel, and further researches the simulation method based on digital radio frequency memory technology to simulate the multidimensional characteristics of multiple radar targets such as delay, doppler frequency, amplitude and angle by one channel. Mathematical simulation experiments are designed to verify the feasibility of the method. © 2023 SPIE.

关键词： Radio waves

来源：评论

学校读者我要写书评

暂无评论

Multiple Algorithms Against Multiple Hardware Architectures: Data-Driven Exploration on Deep Convolution Neural Network 16th

Multiple Algorithms Against Multiple Hardware Architectures:...

引用

16th IFIP WG 10.3 International Conference on Network and Parallel Computing, NPC 2019

作者： Xu, Chongyang Luan, Zhongzhi Gao, Lan Wang, Rui Zhang, Han Zhang, Lianyi Liu, Yi Qian, Depei Beihang University Beijing China Science and Technology on Special System Simulation Laboratory Beijing Simulation Center Beijing China

ISBN: (纸本)9783030307080

With the rapid development of deep learning (DL), various convolution neural network (CNN) models have been developed. Moreover, to execute different DL workloads efficiently, many accelerators have been proposed. To guide the design of both CNN models and hardware architectures for a high-performance inference system, we choose five types of CNN models and test them on six processors and measure three metrics. With our experiments, we get two observations and conduct two insights for the design of CNN algorithms and hardware architectures. © 2019, IFIP International Federation for Information Processing.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

Research on weapon system portfolio selection based on combat network modeling 11

Research on weapon system portfolio selection based on comba...

引用

11th Annual IEEE International systems Conference, SysCon 2017

作者： Cheng, Cheng Li, Jichao Zhao, Qingsong Jiang, Jiang Yu, Lixin Shang, Huilin College of Information System and Management National University of Defense Technology Changsha China Science and Technology on Special System Simulation Laboratory Beijing Simulation Center Beijing China

ISBN: (纸本)9781509046225

Joint operations and combat system-of-systems encounter have become major developing trends of modern warfare. Weapon system portfolio selection attracts much attention, because it is closely related to the production, deployment and operation of weapons, which is a crucial factor determining the outcome of the war. This paper presents a portfolio selection model of the weapon systems based on the combat network. Firstly, we describe the portfolio selection problem of weapon systems based on the combat network, introduce the concept of operation loop, and analyze difficulties and strategies of the optimization problem. Then on this basis, the combat network of the weapon equipment system is modeled, and the operational capability evaluation index of weapon system-of-systems based on the operational loop is constructed. Portfolio selection model based on combat network is established to maximize the combined combat capability of the weapon as an optimization target, and the capacity demand and the expense restriction as constraints. Finally, we take the missile defense system as a example to demonstrate the whole calculation process and results of the weapon system portfolio selection based on the combat network model. It shows that our proposed method can achieve a very good performance to solve weapon system portfolio selection problems. © 2017 IEEE.

关键词： system of systems

来源：评论

学校读者我要写书评

暂无评论

NASIL: Neural Network Architecture Searching for Incremental Learning in Image Classification 1

引用

11th International Symposium on Parallel Architectures, Algorithms and Programming, PAAP 2020

作者： Fu, Xianya Li, Wenrui Chen, Qiurui Zhang, Lianyi Yang, Kai Qing, Duzheng Wang, Rui Beihang University Beijing China Science and Technology on Special System Simulation Laboratory Beijing Simulation Center Beijing100854 China

ISBN: (数字)9789811600104

ISBN: (纸本)9789811600098

"Catastrophic forgetting" and scalability of tasks are two major challenges of incremental learning. Both of these issues were related to the insufficient capacity of machine learning model and the insufficiently trained weights as the increasing of tasks. In this paper, we try to figure out the impact of the neural network architecture to the performance of incremental learning in the case of image classification. During the increasing of tasks, we propose to use neural network architecture searching (NAS) to find a structure that fits the new tasks collection better. We build a NAS environment with reinforcement learning as the searching strategy and Long Short-Term Memory network as the controller network. Computation operation and connecting previous nodes are selected for each layer in the search phase. For each time a new group of tasks is added, the neural network architecture is searched and reorganized according to the training data set. To speed up the searching, we design a parameter sharing mechanism, in which the same building blocks in each layer share a group of parameters. We also introduce the quantified-parameter building blocks into the NAS, to identify the best candidate during each round of searching. We test our solution in cifar100 data set, the average accuracy outperforms the current representative solutions (LwEMC, iCaRL, GANIL) by 24.92%, 5.62%, and 3.6%, respectively, the more tasks added, the better our solution performs. © 2021, Springer Nature Singapore Pte Ltd.

关键词： Network architecture

来源：评论

学校读者我要写书评

暂无评论

Pin-Tool based execution backtracking 13th

Pin-Tool based execution backtracking

引用

13th Conference on Advanced Computer Architecture, ACA 2020

作者： Wei, Shuangjian Ji, Weixing Chen, Qiurui Wang, Yizhuo Beijing Institute of Technology Beijing100081 China Science and Technology on Special System Simulation Laboratory Beijing Simulation Center Beijing100854 China

ISBN: (纸本)9789811581342

Checkpoint/restart is a common fault tolerant technique which periodically dump state to reliable storage and restart applications after failure. Most of existing checkpoint/restart implementations only handle volatile state and lack of support for persistence state of applications. Even the algorithm specifically designed for file checkpointing may not support complex operations and some need to modify source code. This paper presents a new checkpoint technique, which use dynamic instrumentation to temporarily cache disk operations in memory, and use existing memory checkpoint tool to dump or restore process state at runtime. We show that not only can this method create regular checkpoints for both volatile and persistence state, but also has important applications in execution backtracking. © Springer Nature Singapore Pte Ltd 2020.

关键词： Cache memory

来源：评论

学校读者我要写书评

暂无评论

A generation algorithm of Latin hypercube sampling based on original sampling points 27

A generation algorithm of Latin hypercube sampling based on ...

引用

27th European Modeling and simulation Symposium, EMSS 2015

作者： Liu, Zhizhao Li, Wei Yang, Ming Xu, Jun Qing, Duzheng Control and Simulation Center Harbin Institute of Technology Harbin150080 China Science and Technology on Special System Simulation Laboratory Beijing Simulation Center Beijing100854 China

ISBN: (纸本)9788897999485

In order to utilize original sampling points as many as possible to construct a Latin Hypercube Sampling (LHS), a generation algorithm of LHS is proposed. The original sampling can be an arbitrary sampling and the new sampling is a LHS strictly. As the high structure of LHS, only part of original sampling points can be used for LHS. There are several problems including how to select the reserved original sampling points and generate the new ones if necessary. Matrices are used to describe the original sampling points and the original sampling points are reserved based on an adjacent matrix. The new sampling points are generated by a subtraction rule. Numerical experiments show that the generation algorithm is effective and saves a lot of running time for simulation systems.

关键词： Geometry

来源：评论

学校读者我要写书评

暂无评论

Full Polarization Radar Target Recognition Based on Ensemble Learning Algorithm 1

引用

19th Chinese Intelligent systems Conference, CISC 2023

作者： Yang, Zongkai Zhao, Jingcheng Li, Yanhan Xuan, Shiyang Zhang, Ke Song, Youbin Beihang University Beijing100083 China Science and Technology on Special System Simulation Laboratory Beijing Simulation Center Beijing100854 China

ISBN: (数字)9789819968862

ISBN: (纸本)9789819968855

The polarization characteristic of radar targets is a significant research object in the fields of radar information processing and electronic countermeasures. The study aims to identify the targets by polarization information and machine learning. Points on the Poincaré sphere represent the polarization states. With the change of incident angle and frequency, polarization properties are mapped onto the curve of the Poincaré sphere. Combination of the individual classifiers in machine learning generated the final learning model with a determined strategy. The ensemble learning classifier recognized targets by analyzing the difference of polarization curve. The results indicate that the success rate of recognizing five aircraft targets up to 95.8 %. Full polarization information and ensemble learning algorithm is effective for classification of targets. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Polarization

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：