检索结果-内蒙古大学图书馆

24th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2024

作者： Cao, Yongxiang Jiang, Jixiang Zhao, Guocheng Wang, Wei Jiang, Hongxu Song, Yanfei School of Computer Science and Engineering Beihang University Beijing China Science and Technology on Complex System Control and Intelligent Agent Cooperation Laboratory Beijing China

ISBN: (纸本)9789819615445

The systolic array (SA) architecture is widely used in accelerator/AI chip design due to its excellent matrix multiplication acceleration effect. However, the mismatch between the traditional sparse matrix compression format and the SA architecture results in poor SA performance in computing sparse matrices. To solve the problem, we propose a method for co-designing the sparse compression format and SA architecture. This work proposes the Vector Group Compressed Coordinate (VGCC) data stream compression format and designs a sparse SA architecture based on hardware encoding and decoding. VGCC can handle data streams of any size without increasing the index bit width. The VGCC compression size is significantly smaller than general sparse compression formats such as COO/CSR. In a sparse matrix with a sparsity of 90%, it saves 67.13% of storage space compared to the state-of-the-art ECOO data stream compression format and achieves an average acceleration of 1.54 times. We also propose a VGCC-friendly SA architecture. Our SA architecture engine reduces logic consumption by 51.6% compared to Sparse TPU while achieving a 3.23x speedup. To mitigate the issue of high no-load rates in systolic arrays, we introduce a Data Transmission Priority (DTP) data matching algorithm. With sparsity levels ranging from 50% to 90%, the algorithm’s calculation speed is enhanced by 1.124x to 1.991x. We implemented a sparse DNN accelerator on an FPGA Zynq UltraScale+MPSoC ZCU102, utilizing the VGCC compression format proposed in this study for compressing input data. When compared to the S2 Engine, the accelerator achieved a 3.37x speed improvement in matrix multiplication with 75% sparsity, significantly enhancing the speed of sparse matrix multiplication. The accelerator achieved an average throughput of 1079.33GOP/s when inferencing DNNs. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Systolic arrays

来源：评论

学校读者我要写书评

暂无评论

BOAformer: Object Detection Network Based on Background-Object Attention Transformer 6

BOAformer: Object Detection Network Based on Background-Obje...

引用

6th International Conference on Communications, Information System and computer engineering, CISCE 2024

作者： Yang, Yong Liang, Chenbin Huang, Shuying Wang, Xiaozheng School of Computer Science and Technology Tiangong University Tianjin China School of Electronics and Information Engineering Tiangong University Tianjin China School of Software Tiangong University Tianjin China School of Control Science and Engineering Tiangong University Tianjin China

ISBN: (纸本)9798350353174

At present, deep learning has achieved great success in the field of object detection. To ensure that positive samples in the image are not missed, most deep-learning object detection methods set many prediction boxes and use the same prediction operation on them. Although these methods can obtain higher predictive values for high certainty objects, they may identify objects with lower certainty, such as small or fuzzy objects, as negative samples or miss them. To address this issue, this paper presents an object detection network based on background-object attention transformer (BOAformer), which improves the accuracy of object detection by establishing the relationship between background and object features. Firstly, the existing classification model is employed to obtain scores of object prediction boxes. On this basis, a scoring mechanism is designed to filter out high scoring object prediction boxes to reduce the computational complexity of the later network. Then, to improve the accuracy of object detection, BOAformer is constructed, which obtains auxiliary classification values of objects by learning the relationship between background and object features, and adds them to the initial detection results. Finally, many experiments are conducted to verify the effectiveness of BOAformer. The experimental results show that the designed network improves the accuracy of object detection and can detect small or lost objects in the background. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

SCPSN: Spectral Clustering-based Pyramid Super-resolution Network for Hyperspectral Images 24

SCPSN: Spectral Clustering-based Pyramid Super-resolution Ne...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Yang, Yong Zhao, Aoqi Huang, Shuying Wang, Xiaozheng Fan, Yajing School of Computer Science and Technology Tiangong University Tianjin China School of Software Tiangong University Tianjin China School of Control Science and Engineering Tiangong University Tianjin China School of Electronics and Information Engineering Tiangong University Tianjin China

ISBN: (纸本)9798400706868

Single hyperspectral image super-resolution aims to reconstruct a high-resolution hyperspectral image (HRHSI) from an observed low resolution hyperspectral image (LRHSI). Most current methods combine CNN and Transformer structures to directly extract features of all channels in LRHSI for image reconstruction, but they do not consider the interference of redundant information in adjacent bands, resulting in spectral and spatial distortions in the reconstruction results and an increase in model computational complexity. To address this issue, this paper proposes a spectral clustering-based pyramid super-resolution network (SCPSN) to progressively reconstruct HRHSI at different scales. In each image reconstruction layer, a clustering super-resolution block (CSRB) consisting of spectral clustering block (SCB), patch non local attention block (PNAB), and dynamic fusion block (DFB) is designed to achieve the reconstruction of detail features. Specifically, for the high correlation between adjacent spectral bands in LRHSI, a SCB is first constructed to achieve clustering of spectral channels and filtering of hyperchannels. This can reduce the interference of redundant spectral information and the computational complexity of the model. Then, by utilizing the non-local similarity of features within the channel, a patch non-local attention block (PNAB) is constructed to enhance the features of hyperchannels. Next, a dynamic fusion block (DFB) is designed to reconstruct the features of all channels in LRHSI by establishing correlations between enhanced hyperchannels and other channels. Finally, the reconstructed channels are upsampled and added to the corresponding channels to obtain the reconstructed HRHSI. Extensive experiments validate that the performance of SCPSN is superior to that of some other state-of-the-art (SOTA) HSSR methods in terms of visual effects and quantitative metrics. In addition, our model does not require training on large-scale datasets compared to oth

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Probabilistic Tube-based control Synthesis of Stochastic Multi-Agent Systems under Signal Temporal Logic 63

Probabilistic Tube-based Control Synthesis of Stochastic Mul...

引用

63rd IEEE Conference on Decision and control, CDC 2024

作者： Vlahakis, Eleftherios E. Lindemann, Lars Sopasakis, Pantelis Dimarogonas, Dimos V. Kth Royal Institute of Technology Division of Decision and Control Systems School of Electrical Engineering and Computer Science Stockholm10044 Sweden University of Southern California Viterbi School of Engineering Department of Computer Science Los AngelesCA90089 United States Queen's University School of Electronics Electrical Engineering and Computer Science Northern Ireland Belfast United Kingdom

ISBN: (纸本)9798350316339

We consider the control design of stochastic discrete-time linear multi-agent systems (MASs) under a global signal temporal logic (STL) specification to be satisfied at a predefined probability. By decomposing the dynamics into deterministic and error components, we construct a probabilistic reachable tube (PRT) as the Cartesian product of reachable sets of the individual error systems driven by disturbances lying in confidence regions (CRs) with a fixed probability. By bounding the PRT probability with the specification probability, we tighten all state constraints induced by the STL specification by solving tractable optimization problems over segments of the PRT, and relax the underlying stochastic problem with a deterministic one. This approach reduces conservatism compared to tightening guided by the STL structure. Additionally, we propose a recursively feasible algorithm to attack the resulting problem by decomposing it into agent-level subproblems, which are solved iteratively according to a scheduling policy. We demonstrate our method on a ten-agent system, where existing approaches are impractical. © 2024 IEEE.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

Multi-Agent Imitation Learning Based Energy Management of a Microgrid With Hybrid Energy Storage and Real-Time Pricing

引用

IEEE Internet of Things Journal 2025年第12期12卷 19801-19817页

作者： Gao, Shuhua Xu, Yizhuo Zhang, Zhaoqian Wang, Zhengfang Zhou, Xiaoyu Wang, Jing Shandong University School of Control Science and Engineering Jinan250100 China The Hong Kong Polytechnic University Department of Electronic and Information Engineering Hong Kong Hong Kong National University of Singapore Department of Electrical and Computer Engineering Singapore117583 Singapore Second Construction Limited Company of China Construction Eighth Engineering Division Jinan Shandong China

Microgrids equipped with hybrid energy storage systems (ESSs) are increasingly critical for balancing the intermittency of renewable energy sources (RESs) and the fluctuations in demand. This paper introduces a novel Multi-Agent Imitation Learning (MAIL) framework for real-time energy management in microgrids, particularly under real-time pricing conditions. The approach leverages a problem decomposition strategy, which separates the energy management task into two phases: first, optimal actions for each ESS are estimated using individual agents, each employing a deep neural network (DNN) trained to emulate an ideal mixed-integer linear programming (MILP) solver;then, a one-step online optimization reacts to these decisions to optimize the remaining system components holistically. Rigorous proofs establish the equivalence of the decomposed subproblems to the original optimization problem, ensuring integrity and effectiveness. Comparative simulation studies with real-world data reveal that our MAIL framework offers significant cost advantages and superior training efficiency relative to both traditional single and emerging multi-agent reinforcement learning (RL) methods. Notably, the operational costs achieved through our approach are markedly lower than those of its counterparts, closely approximating the theoretical minimum. This underscores the method's proficiency and its potential as an effective solution for microgrid energy management. Moreover, the MAIL strategy's multi-agent design allows for straightforward scalability to more extensive systems involving multiple ESSs and shows promising potential for cooperative management across interconnected microgrids. © 2014 IEEE.

关键词： Costs

来源：评论

学校读者我要写书评

暂无评论

Analysis of the Direction of Library Development in the Future Based on Knowledge Graph Analysis 6th

Analysis of the Direction of Library Development in the Futu...

引用

6th International Conference on Innovative Computing, IC 2023

作者： Dong, Jie Luo, Lingxiang School of Electrical and Control Engineering Shenyang Jianzhu University Liaoning Shenyang China School of Computer Science and Engineering Shenyang Jianzhu University Liaoning Shenyang China

ISBN: (纸本)9789819920914

Visual analysis of library reader service research and linear regression analysis are conducted to point out the closeness and changing rules of different influencing factors and the overall service quality of the library, providing suggestions for the optimization direction of library service problems in the future. In the CNKI database, the data in the field of "library reader service" is used as the initial data, and the knowledge map is used to analyze and study the research hotspots in this field. The questionnaire is designed based on the results, and the regression analysis is conducted on the questionnaire data. The research hotspots in the field of library reader service mainly focus on four aspects: reader service, reading service, environmental service and activity organization. By combining these four aspects to provide multi-level services, libraries can adapt to the rapid development of libraries in the new century. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Regression analysis

来源：评论

学校读者我要写书评

暂无评论

A Method for Constructing an Olfactory Dataset Based on fNIRS 9

A Method for Constructing an Olfactory Dataset Based on fNIR...

引用

9th International Conference on Systems and Informatics, ICSAI 2023

作者： Li, Wei He, Haobo Li, Guang Kuang, Lidan Zhang, Jin Changsha University of Science and Technology School of Computer and Communication Engineering Changsha410076 China Zhejiang University School of Control Science and Engineering Hangzhou310027 China

ISBN: (纸本)9798350383706

Biomimetic olfaction models have been applied to various fields such as image recognition, text classification, electroencephalography (EEG) recognition, and spatio-temporal sequence prediction. However, the challenge of obtaining olfactory data complicates the construction of olfactory datasets, ultimately resulting in a scarcity of applications for biomimetic olfaction models in olfactory tasks. In this work, we begin by conducting a comprehensive comparison of existing biomimetic olfactory models from various aspects such as structure, degree of biomimicry, and performance. Following this, prefrontal lobe olfactory data from ten subjects were collected for three different odors using functional Near-Infrared Spectroscopy (fNIRS). Finally, the data underwent denoising through the use of various data processing methods, and the quantity of the olfactory dataset was expanded. The olfactory dataset constructed in this study can be used to evaluate the performance of biomimetic olfactory models in processing olfactory tasks, which may aid in advancing the development of these models in real-world olfactory applications. © 2023 IEEE.

关键词： Image recognition

来源：评论

学校读者我要写书评

暂无评论

EVADE: Targeted Adversarial False Data Injection Attacks for State Estimation in Smart Grid

IEEE Transactions on Sustainable Computing

引用

IEEE Transactions on Sustainable Computing 2024年第3期10卷 534-546页

作者： Tian, Jiwei Shen, Chao Wang, Buhong Ren, Chao Xia, Xiaofang Dong, Runze Cheng, Tianhao Ministry of Education Key Laboratory for Intelligent Networks and Network Security School of Cyber Science and Engineering Xi'an Jiaotong University Xi'an China Air Traffic Control and Navigation College Air Force Engineering University Xi'an China Information and Navigation College Air Force Engineering University Xi'an China School of Electrical and Electronic Engineering Nanyang Technological University Singapore School of Computer Science and Technology Xidian University Xi'an China

Although conventional false data injection attacks can circumvent the detection of bad data detection (BDD) in sustainable power grid cyber physical systems, they are easily detected by well-trained deep learning-based detectors. Still, state estimation models with deep leaning-based detectors are not secure due to the vulnerabilities and fragility of deep learning models. Using the related laws of conventional false data injection attacks and adversarial sample attacks, this paper proposes the targEted adVersarial fAlse Data injEction (EVADE) strategy to explore targeted adversarial false data injection attacks for state estimation in Smart Grid. The proposed EVADE attack strategy selects key state variables based on adversarial saliency maps to improve the attack efficiency and perturbs as few state variables as possible to reduce the attack cost. In this way, the EVADE attack strategy can bypass the detection of BDD and neural attack detection (NAD) methods (that is, maintaining deep stealthy) with a high success rate and achieve the attack target simultaneously. Experimental results demonstrate the effectiveness of the proposed strategy, posing serious and pressing concerns for sustainable cyber physical power system security. © 2016 IEEE.

关键词： Cyber attacks

来源：评论

学校读者我要写书评

暂无评论

EEG-based comparative measurement of visual fatigue caused by stereoscopic depth perception

EEG-based comparative measurement of visual fatigue caused b...

引用

The 2023 International Conference on Display Technology, (ICDT 2023)

作者： Jia, Lixiu Huang, Sheng Ding, Yiquan Feng, Lihang Huang, Xiaohua School of Computer Engineering Nanjing Institute of Technology Jiangsu Nanjing China School of Electrical Engineering and Control Science Nanjing Tech University Jiangsu Nanjing China

In this study, we investigated the influence of stereoscopic depth on visual fatigue based on subjective and EEG measurement. 19 channels of EEG data of 17 subjects before and after watching 3D movie with different depth ranges were recorded, and the power of δ, θ, α, β, power spectral parameters (gravity frequency, power spectral entropy) were calculated and analyzed. The results showed that visual fatigue caused by 3D enhanced movies with larger negative disparity was more significantly severe than that of 3D normal movies. © 2023, John Wiley and Sons Inc. All rights reserved.

关键词： Stereo image processing

来源：评论

学校读者我要写书评

暂无评论

Research on Optimization Method of Visual Odometer Based on Point Line Feature Fusion 23

Research on Optimization Method of Visual Odometer Based on ...

引用

7th International Conference on High Performance Compilation, Computing and Communications, HP3C 2023

作者： Zhang, Guoyin Yuan, Quande Liu, Yifan School of Information and Control Engineering Jilin Institute of Chemical Technology Jilin China School of Computer Technology and Engineering Changchun Institute of Technology Jilin Changchun China

ISBN: (纸本)9781450399883

In low-texture-scenes, point feature based simultaneous location and map (SLAM) building algorithms are difficult to extract enough valid feature points, so that the SLAM system cannot work properly. To solve such problems, this paper proposes a visual SLAM algorithm for fusion of point and line features. On the basis of ORB point feature extraction, the LSD line feature algorithm is introduced and the extracted line segments are represented by the Plücker coordinate method. In the online feature matching stage, the LBD line segment descriptor is introduced, and the quadtree algorithm is used to avoid the concentration problem of point feature extraction. The progressive sampling uniformity (PROSAC) algorithm and geometric constraints are used to filter out incorrect matching of point line features, thereby improving positioning accuracy. Then, in the process of feature fusion, the idea of point line feature weighting is introduced to construct a new error function for point line fusion, minimizing the error function to obtain more accurate positioning information. Finally, the publicly available dataset was used to conduct experiments. © 2023 ACM.

关键词： Feature extraction

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：