检索结果-内蒙古大学图书馆

Chinese Automation Congress (CAC)

作者： Shuang Liu Zhicheng Liu Yuhai Li Wancheng Liu Chenyang Ge Longjun Liu Institute of Artificial Intelligence and Robotics (IAIR) Xi’an Jiaotong University Xi’an China Science and Technology on Electro-Optical Information Security Control Laboratory Tianjin China

most YOLO object detection neural networks prefer to focus on traditional RGB image, but previous studies rarely consider special YOLO network with compact architecture for infrared image. In this paper, we analyze original YOLO network architecture, and we propose a compact YOLO based network by using different blocks from previous work for small target detection on infrared image. We use small target detection layer, GhostConv convolution, Focus structure, Focal EIOU Loss and soft NMS modules to improve original YOLO network structure, which improves the accuracy and speed of target detection for infrared image. The experimental results show that the accuracy of target detection of original model can be effectively improved by adding the small target detection layer, the Focus structure and the BI-FPN structure. By replacing the GhostConv convolution, the calculation speed is significantly improved. At last, after the model is quantized and deployed with NCNN framework, the object detection speed can reached to 35ms per infrared image.

关键词： Automation Convolution Neural networks Object detection Network architecture

来源：评论

学校读者我要写书评

暂无评论

Modulation Recognition Based on Lightweight Residual Network via Binary Quantization

Modulation Recognition Based on Lightweight Residual Network...

引用

International Conference on Computer and communications (ICCC)

作者： Heng Ji Wangyang Xu Lu Gan Zhengwu Xu Science and Technology on Communication Information Security Control Laboratory Jiaxing China University of Electronic Science and Technology of China Chengdu China Yibin Institute of UESTC Yibin China

ISBN: (纸本)9781665409513

Signal modulation recognition is an important and complicated problem in spectrum monitoring. In recent years, many advanced deep learning methods have been proposed for this problem. However, the high computational cost of deep neural networks largely hinders practical deployment. In this paper, we consider the signal modulation recognition problem via the deep residual network and propose a binary quantization scheme to reduce the model size and memory cost. This is achieved by simultaneously increasing shortcuts and maintaining full-precision weights and outputs of key layers in binary residual network in a simple way. The results show that with competitive accuracy, 95% of multiply–accumulate operations of model are converted into binary operations under 0 dB snr. In summary, the proposed method can effectively reduce the parameters and complexity of recognition model.

关键词： Deep learning Quantization (signal) Computational modeling Simulation Neural networks Modulation Robustness

来源：评论

学校读者我要写书评

暂无评论

Phase noise model for continuous-variable quantum key distribution using a local local oscillator

引用

Physical Review A 2021年第3期104卷 032608-032608页

作者： Yun Shao Heng Wang Yaodi Pi Wei Huang Yang Li Jinlu Liu Jie Yang Yichen Zhang Bingjie Xu Science and Technology on Communication Security Laboratory Institute of Southwestern Communication Chengdu 610041 China State Key Laboratory of Information Photonics and Optical Communications Beijing University of Posts and Telecommunications Beijing 100876 China

The value of residual phase noise, after phase compensation, is one of the key limitations of performance improvement for continuous-variable quantum key distribution using a local local oscillator (LLO CV-QKD) system, since it is the major excess noise. However, due to the nonideality of the realistic devices implemented in practice, for example, imperfect lasers, detectors, and unbalanced interferometers, the value of residual phase noise in the current system is still relatively large. Here, we develop a phase noise model to improve the phase noise tolerance of the LLO CV-QKD schemes. In our model, part of the phase-reference measurement noise associated with detection efficiency and electronic noise of Bob's detector as well as a real-time monitored phase-reference intensity at Bob's side is considered trusted because it can be locally calibrated by Bob. We show that using our phase noise model can significantly improve the secure key rate and transmission distance of the LLO CV-QKD system. We further conduct an experiment to substantiate the superiority of the phase noise model. Based on experimental data of a LLO CV-QKD system in the 25-km optical fiber channel, we demonstrate that the secure key rate under our phase noise model is approximately 40% higher than that under the conventional phase noise model.

关键词： Quantum cryptography

来源：评论

学校读者我要写书评

暂无评论

Modulation Recognition Based on Lightweight Residual Network via Hybrid Pruning

Modulation Recognition Based on Lightweight Residual Network...

引用

International Conference on Computer and communications (ICCC)

ISBN: (纸本)9781665409513

Signal modulation recognition is an important and complicated problem in spectrum monitoring. In recent years, many advanced deep learning methods have been proposed for this problem. However, the high computational cost of deep neural networks largely hinders practical deployment. In this paper, we consider the signal modulation recognition problem via the deep residual network and propose a hybrid Pruning scheme to reduce the model size and memory cost. This is achieved by simultaneously enforcing channel-level and layer-level sparsity in residual network in a simple way. We take wide and large networks as input models, but during training, insignificant channels and layers are automatically identified and pruned afterwards, yielding compact and shallower networks with comparable accuracy. The results show that with 0 dB snr, the trainable parameters and multiply–accumulate operations are both reduced by 2/3. In summary, the proposed method can effectively reduce the parameters and complexity of recognition model.

关键词： Deep learning Training Computational modeling Simulation Neural networks Modulation Robustness

来源：评论

学校读者我要写书评

暂无评论

Direction Modulation Design for UAV Assisted by IRS with discrete phase shift

arXiv

引用

arXiv 2024年

作者： Li, Maolin Gao, Wei Wu, Qi Shu, Feng Pan, Cunhua Wu, Di The School of Information and Communication Engineering Hainan University Haikou570228 China The School of Information and Communication Engineering Collaborative Innovation Center of Information Technology Hainan University Haikou570228 China The School of Electronic and Optical Engineering Nanjing University of Science and Technology Nanjing210094 China The National Mobile Communications Research Laboratory Southeast University China The School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai200240 China The State Key Laboratory of Rail Traffic Control and Safety Beijing Jiaotong University Beijing100044 China

As a physical layer security technology, directional modulation (DM) can be combined with intelligent reflect-ing surface (IRS) to improve the security of drone communications. In this paper, a directional modulation scheme assisted by the IRS is proposed to maximize the transmission rate of unmanned aerial vehicle (UAV) secure communication. Specifically, with the assistance of the IRS, the UAV transmits legitimate information and main-tains its constellation pattern at the location of legitimate users on the ground, while the constellation pattern is disrupted at the eavesdropper's location. In order to solve the joint optimization problem of digital weight coefficients, UAV position, and IRS discrete phase shift, firstly, the digital weight vector and UAV position are optimized through power minimization. Secondly, three methods are proposed to optimize IRS phase shift, namely vector trajectory (VT) method, cross entropy vector trajectory (CE-VT) algorithm, and block coordinate descent vector trajectory (BCD-VT) algorithm. Compared to traditional cross entropy (CE) methods and block coordinate descent (BCD) methods, the proposed CE-VT and BCD-VT algorithms can improve transmission rate performance. The numerical results validate the effectiveness of the optimization scheme in IRS assisted UAV communication. Copyright © 2024, The Authors. All rights reserved.

关键词： Vectors

来源：评论

学校读者我要写书评

暂无评论

Message from the General Conference Chairs

Proceedings - 2021 International Conference on Networking an...

引用

Proceedings - 2021 International Conference on Networking and Network Applications, NaNA 2021 2021年 XIX页

作者： Wang, Zhenjiang Gu, Naijie Ho, Pin-Han Chen, Shunyang Lijiang Culture and Tourism College China University of Science and Technology of China China University of Waterloo Canada Science and Technology on Communication Information Security Control Laboratory China

来源：评论

学校读者我要写书评

暂无评论

FPGA-based Acceleration of Structured Light Depth Estimation

FPGA-based Acceleration of Structured Light Depth Estimation

引用

Chinese Automation Congress (CAC)

作者： Shijie Wang Xiaodong Deng Wancheng Liu Yuhai Li Shitao Chen Longjun Liu Institute of Artificial Intelligence and Robotics (IAIR) Xi’an Jiaotong University Xi’an China Science and Technology on Electro-Optical Information Security Control Laboratory Tianjin China

Stereo depth estimation is widely applied in computer vision and intelligent robotic systems. For real-time mobile stereo vision applications, how to efficiently design and implement stereo matching on resource-limited hardware platforms is still a grand challenge. In this paper, we propose and implement a low-power, high-resolution stereo matching hardware architecture on field-programmable gate array (FPGA) with the input of structured light camera. In order to obtain high-quality point cloud data, we present an image preprocessing module and an improved cost function for stereo matching algorithm. Furthermore, we leverage hardware description language to implement our stereo matching computing architecture and all algorithm models are evaluated on the Xilinx ZCU102 development board, which only takes about 61147 LUTs and 786 BRAMs. Compared with high-performance CPUs, the computing speed is significantly improved, 1.84 times faster than Intel i7-10700@2.90GHz, and the power consumption is only 3.915W.

关键词： Costs Power demand Robot vision systems Estimation Computer architecture Hardware Real-time systems

来源：评论

学校读者我要写书评

暂无评论

Design an Efficient DNN Inference Framework with PS-PL Synergies in FPGA for Edge Computing

Design an Efficient DNN Inference Framework with PS-PL Syner...

引用

Chinese Automation Congress (CAC)

作者： Wancheng Liu Yuhai Li Yining Yang Jinyu Zhu Longjun Liu Science and Technology on Electro-Optical Information Security Control Laboratory Tianjin China Institute of Artificial Intelligence and Robotics (IAIR) Xi’an Jiaotong University Xi’an China

in recent years, Deep Neural Network (DNN) based methods have achieved great success for changing inflexible machine to intelligent and live system. Internet of Things (IoT) applications equipped with DNN in domains such as self-driving, speech processing and video surveillance particular challenges. Since DNNs are memory-/computing- intensive applications, FPGAs are customized to accelerate DNN on edge system due to their low latency and high energy efficiency. However, embedded CPU is usually essential for IoT system and single accelerator of FPGA is hard to gain maximized performance of the whole IoT system. Therefore, we design an efficient DNN inference framework with programming system (PS, e.g., ARM CPU) and programming logic (PL, e.g. FPGA) synergies for intelligent IoT system. Compared with the common IoT system only separately leverages FPGA for DNN accelerating and ARM for other tasks, our framework can make ARM and FPGA jointly accelerate DNN inference from the beginning. Our PS-PL co-design approach can take full advantage of the merits of ARM (control operations) and FPGA (parallel computing operations) for various DNN model accelerating. We implemented our co-design framework on Zynq SoCs. The experimental results show that our framework can achieve maximal 205.9 GOPs/s DNN inference accelerating under 6.2W while it would not affect other CPU tasks running on IoT system.

关键词： Program processors Computational modeling Neural networks Programming Parallel processing Video surveillance Internet of Things

来源：评论

学校读者我要写书评

暂无评论

Factored Trace Lasso Based Linear Regression Methods: Optimizations and Applications 5th

Factored Trace Lasso Based Linear Regression Methods: Optimi...

引用

5th International Conference on Cognitive Systems and Signal Processing, ICCSIP 2020

作者： Zhang, Hengmin Du, Wenli Liu, Xiaoqian Zhang, Bob Qian, Feng School of Information Science and Engineering Key Laboratory of Advanced Control and Optimization for Chemical Processes Ministry of Education East China University of Science and Technology Shanghai200237 China Department of Computer Information and Cyber Security Jiangsu Police Institute Nanjing210031 China Department of Computer and Information Science University of Macau 999078 China

ISBN: (纸本)9789811623356

Consider that matrix trace lasso regularized convex p -norm with p= 1, 2 regression methods usually have the higher computational complexity due to the singular value decomposition (SVD) of larger size matrix in big data and information processing. By factoring the matrix trace lasso into the squared sum of two Frobenius-norm, this work studies the solutions of both adaptive sparse representation (ASR) and correlation adaptive subspace segmentation (CASS), respectively. Meanwhile, the derived models involve multi-variable nonconvex functions with at least two equality constraints. To solve them efficiently, we devise the nonconvex alternating direction multiplier methods (NADMM) with convergence analysis satisfying the Karush-Kuhn-Tucher (KKT) conditions. Finally, numerical experiments to the subspace clustering can show the less timing consumptions than CASS and the nearby performance of our proposed method when compared with the existing segmentation methods like SSC, LRR, LSR and CASS. © 2021, Springer Nature Singapore Pte Ltd.

关键词： Singular value decomposition

来源：评论

学校读者我要写书评

暂无评论

Twitter Bot Detection with Multi-Head Attention and Supervised Contrastive Learning

Twitter Bot Detection with Multi-Head Attention and Supervis...

引用

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

作者： Liu Qiang Lu Jiazhong Huang Yuanyuan Zhang Weisha Lv Jun School of Cybersecurity (Xin Gu Industrial College) Chengdu University of Information Technology Chengdu China Advanced Cryptography and System Security Key Laboratory of Sichuan Province Chengdu China SUGON Industrial Control and Security Center Chengdu China School of Foreign Languages University of Electronic Science and Technology of China Chengdu China MetaX Integrated Circuits (Shanghai) Co. Ltd

ISBN: (数字)9798331519254

ISBN: (纸本)9798331519261

The task of detecting bots on Twitter is an indispensable task for combating social network disinformation, curbing rumors, and managing online public opinion. However, today's social bots often disguise themselves by mimicking various aspects of real user characteristics to escape detection, which poses an unprecedented challenge to the existing Twitter bot detection task. To address this, an innovative detection method is presented in this study that improves the accuracy of Twitter bot detection by realizing the effective fusion of multi-dimensional features through the multi-head attention mechanism and applying it to a relational graph convolutional network. It also combines supervised contrastive learning to strengthen the differentiation ability of the high-latitude features. Experiments conducted on the publicly available dataset Twibot-20, compared to the baseline, the proposed method achieves approximately 1% better results in both Accuracy and the Matthews Correlation Coefficient, proving its effectiveness.

关键词： Training Attention mechanisms Accuracy Social networking (online) Graph convolutional networks Blogs Contrastive learning information processing Feature extraction Fake news

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：