检索结果-内蒙古大学图书馆

arXiv 2021年

作者： Ye, Haishan Zhang, Tong Shenzhen Research Institute of Big Data Chinese University of Hong Kong Shenzhen Hong Kong Hong Kong University of Science and Technology Hong Kong

Due to the rapid growth of smart agents such as weakly connected computational nodes and sensors, developing decentralized algorithms that can perform computations on local agents becomes a major research direction. This paper considers the problem of decentralized Principal components analysis (PCA), which is a statistical method widely used for data analysis. We introduce a technique called subspace tracking to reduce the communication cost, and apply it to power iterations. This leads to a decentralized PCA algorithm called DeEPCA, which has a convergence rate similar to that of the centralized PCA, while achieving the best communication complexity among existing decentralized PCA algorithms. DeEPCA is the first decentralized PCA algorithm with the number of communication rounds for each power iteration independent of target precision. Compared to existing algorithms, the proposed method is easier to tune in practice, with an improved overall communication cost. Our experiments validate the advantages of DeEPCA empirically. © 2021, CC BY.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Heterogeneous Parallel Computing Based Real-Time Chaotic Video Encryption and its Application to Drone-Oriented Secure Communication

SSRN

引用

SSRN 2023年

作者： Jiang, Dong Shi, Fan-Feng Li, Tao Li, Yi-Fei Shan, Dan School of Information Engineering Yangzhou Polytechnic Institute Yangzhou225007 China School of Internet Anhui University Hefei230039 China College of Physical Science and Technology Yangzhou University Yangzhou225009 China National Engineering Research Center of Agro-Ecological Big Data Analysis and Application Anhui University Hefei230601 China

This paper proposes a real-time video encryption protocol based on multi-round confusion-diffusion architecture and heterogeneous parallel computing. It leverages the powerful computing capacity of Central Processing Unit (CPU) and the high parallel capability of Graphics Processing Unit (GPU) to perform byte generation, confusion and diffusion operations concurrently, thereby improving computational efficiency. Statistical and security analysis show that the proposed protocol exhibits exceptional statistical properties and provides resistance against different types of attacks. Encryption speed evaluation demonstrates that it can realize latency-free 640x640 30FPS video encryption using Intel Core i7-8700 and NVIDIA GeForce GTX 1060, with an average encryption time of 20.43ms, despite performing seven rounds of confusion and three rounds of diffusion operations on each frame. Additionally, the proposed protocol is adopted to implement a drone-oriented secure video communication system, with NVIDIA Jetson Xavier NX (NVIDIA Camel ARM CPU and Volta GPU), achieving latency-free 256x256 29FPS video encryption. © 2023, The Authors. All rights reserved.

关键词： Program processors

来源：评论

学校读者我要写书评

暂无评论

Understanding Large Language Model Supply Chain: Structure, Domain, and Vulnerabilities

arXiv

引用

arXiv 2025年

作者： Hu, Yanzhe Wang, Shenao Nie, Tianyuan Zhao, Yanjie Wang, Haoyu Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Huazhong University of Science and Technology China

Large Language Models (LLMs) have revolutionized artificial intelligence (AI), driving breakthroughs in natural language understanding, text generation, and autonomous systems. However, the rapid growth of LLMs presents significant challenges in the security and reliability of the Large Language Model Supply Chain (LLMSC), a complex network of open-source components, libraries, and tools essential for LLM development and deployment. Despite its critical importance, the LLMSC remains underexplored, particularly regarding its structural characteristics, domain composition, and security vulnerabilities. To address this gap, we conduct the first empirical study of the LLMSC, analyzing a curated dataset of open-source packages from PyPI and NPM across 14 functional domains. We construct a directed dependency graph comprising 15,725 nodes, 10,402 edges, and 180 unique vulnerabilities to investigate the structural characteristics of the LLMSC and analyze how security risks propagate through its dependency network. Our findings reveal that the LLMSC exhibits a "locally dense, globally sparse" topology, with 79.7% of dependency trees containing fewer than 5 nodes, while a few large trees dominate the ecosystem, accounting for 77.66% of all nodes. The graph is characterized by high-degree hubs, with the top 5 most connected nodes averaging 1,282 dependents each. Security analysis shows that critical vulnerabilities propagate to an average of 142.1 nodes at the second layer of dependency trees and peak at 237.8 affected nodes at the third layer. Notably, cascading risks are concentrated in critical hub nodes such as transformers, which directly or indirectly affect over 1,300 downstream packages. These findings provide quantitative insights into the structural and security dynamics of the LLMSC and emphasize the need for targeted mitigation strategies to enhance ecosystem resilience. Copyright © 2025, The Authors. All rights reserved.

关键词： Risk assessment

来源：评论

学校读者我要写书评

暂无评论

MetaGCD: Learning to Continually Learn in Generalized Category Discovery

MetaGCD: Learning to Continually Learn in Generalized Catego...

引用

International Conference on Computer Vision (ICCV)

作者： Yanan Wu Zhixiang Chi Yang Wang Songhe Feng Key Laboratory of Big Data & Artificial Intelligence in Transportation Ministry of Education Beijing Jiaotong University Beijing China School of Computer and Information Technology Beijing Jiaotong University Beijing China Department of Electrical and Computer Engineering University of Toronto Toronto Canada Department of Computer Science and Software Engineering Concordia University Montreal Canada

In this paper, we consider a real-world scenario where a model that is trained on pre-defined classes continually encounters unlabeled data that contains both known and novel classes. The goal is to continually discover novel classes while maintaining the performance in known classes. We name the setting Continual Generalized Category Discovery (C-GCD). Existing methods for novel class discovery cannot directly handle the C-GCD setting due to some unrealistic assumptions, such as the unlabeled data only containing novel classes. Furthermore, they fail to discover novel classes in a continual fashion. In this work, we lift all these assumptions and propose an approach, called MetaGCD, to learn how to incrementally discover with less forgetting. Our proposed method uses a meta-learning framework and leverages the offline labeled data to simulate the testing incremental learning process. A meta-objective is defined to revolve around two conflicting learning objectives to achieve novel class discovery without forgetting. Furthermore, a soft neighborhood-based contrastive network is proposed to discriminate uncorrelated images while attracting correlated images. We build strong baselines and conduct extensive experiments on three widely used benchmarks to demonstrate the superiority of our method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Optimizing Multilevel Thresholding for Image Segmentation Based on Hybrid Swarm Computation Optimization 16th

An Optimizing Multilevel Thresholding for Image Segmentation...

引用

16th International Conference on Intelligent Information Hiding and Multimedia Signal Processing, IIH-MSP 2020 in conjunction with the 13th International Conference on Frontiers of Information technology, Applications and Tools, FITAT 2020

作者： Dao, Thi-Kien Wang, Hong-Jiang Yu, Jie Nguyen, Huu-Quynh Ngo, Truong-Giang Nguyen, Trong-The Fujian Provincial Key Laboratory of Big Data Mining and Applications Fujian University of Technology Fuzhou350118 China College of Mechanical and Automotive Engineering Fujian University of Technology Fuzhou350118 China Faculty of Computer Science and Engineering Thuyloi University 175 Tay Son Dong Da Hanoi Viet Nam Department of Information Technology Haiphong University of Management and Technology Haiphong Viet Nam

ISBN: (纸本)9789813367562

This paper suggests a solution for the image segmentation (IS) problem with the multilevel thresholding based on one of the latest hybrid swarm computation optimization algorithms, particle swarms, and gravitational search (PSGA). The experimental results are comparable with other state-of-the-art algorithms that show that the PSGA on selected images is better than the competitors. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Zero Stability Well Predicts Performance of Convolutional Neural Networks

arXiv

引用

arXiv 2022年

作者： Chen, Liangming Jin, Long Shang, Mingsheng Chongqing Key Laboratory of Big Data and Intelligent Computing Chongqing Institute of Green and Intelligent Technology Chinese Academy of Sciences China Chongqing School University of Chinese Academy of Sciences China School of Information Science and Engineering Lanzhou University China

The question of what kind of convolutional neural network (CNN) structure performs well is fascinating. In this work, we move toward the answer with one more step by connecting zero stability and model performance. Specifically, we found that if a discrete solver of an ordinary differential equation is zero stable, the CNN corresponding to that solver performs well. We first give the interpretation of zero stability in the context of deep learning and then investigate the performance of existing first- and second-order CNNs under different zero-stable circumstances. Based on the preliminary observation, we provide a higher-order discretization to construct CNNs and then propose a zero-stable network (ZeroSNet). To guarantee zero stability of the ZeroSNet, we first deduce a structure that meets consistency conditions and then give a zero stable region of a training-free parameter. By analyzing the roots of a characteristic equation, we theoretically obtain the optimal coefficients of feature maps. Empirically, we present our results from three aspects: We provide extensive empirical evidence of different depth on different datasets to show that the moduli of the characteristic equation's roots are the keys for the performance of CNNs that require historical features;Our experiments show that ZeroSNet outperforms existing CNNs which is based on high-order discretization;ZeroSNets show better robustness against noises on the input. The source code is available at https://***/LongJin-lab/ZeroSNet. Copyright © 2022, The Authors. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

An optimal transport approach for selecting a representative subsample with application in efficient kernel density estimation

arXiv

引用

arXiv 2022年

作者： Zhang, Jingyi Meng, Cheng Yu, Jun Zhang, Mengrui Zhong, Wenxuan Ma, Ping Center for Statistical Science Tsinghua University China Center for Applied Statistics Institute of Statistics and Big Data Renmin University of China China School of Mathematics and Statistics Beijing Institute of Technology China Department of Statistics University of Georgia United States

Subsampling methods aim to select a subsample as a surrogate for the observed sample. Such methods have been used pervasively in large-scale data analytics, active learning, and privacy-preserving analysis in recent decades. Instead of model-based methods, in this paper, we study model-free subsampling methods, which aim to identify a subsample that is not confined by model assumptions. Existing model-free subsampling methods are usually built upon clustering techniques or kernel tricks. Most of these methods suffer from either a large computational burden or a theoretical weakness. In particular, the theoretical weakness is that the empirical distribution of the selected subsample may not necessarily converge to the population distribution. Such computational and theoretical limitations hinder the broad applicability of model-free subsampling methods in practice. We propose a novel model-free subsampling method by utilizing optimal transport techniques. Moreover, we develop an efficient subsampling algorithm that is adaptive to the unknown probability density function. Theoretically, we show the selected subsample can be used for efficient density estimation by deriving the convergence rate for the proposed subsample kernel density estimator. We also provide the optimal bandwidth for the proposed estimator. Numerical studies on synthetic and real-world datasets demonstrate the performance of the proposed method is superior. Copyright © 2022, The Authors. All rights reserved.

关键词： Density functional theory

来源：评论

学校读者我要写书评

暂无评论

Refining Groundwater Storage Loss to Support Irrigated Agriculture by Integrating Mt-Insar, Grace, and Hydraulic Head Measurements

SSRN

引用

SSRN 2024年

作者： Hu, Jiyuan Zhou, Zheng Wang, Jiabei Qin, Fen Wang, Jiayao Zhang, Rong Wu, Wenhao Wang, Lijun College of Geography and Environmental Science Henan University Kaifeng475004 China Key Laboratory of Geospatial Technology for the Middle and Lower Yellow River Regions Henan University Ministry of Education Kaifeng475004 China Henan Water Conservancy Survey. LTD Zhengzhou450008 China Henan Industrial Technology Academy of Spatial-Temporal Big Data Henan University Zhengzhou450046 China Hunan Province Key Laboratory of Coal Resources Clean-utilization and Mine Environment Protection Hunan University of Science and Technology Xiangtan411201 China

Aquifers supporting irrigated agriculture in Henan Plain in China (HNP) are under immense stress due to water scarcity and extensive. To assist in establishing a crop planting structure in this region that aligns more with the capacity of water resources, we have developed a refined methodology for estimating Groundwater Storage Anomalies (GWSA). This method integrates Multi-temporal Interferometric Synthetic Aperture Radar (MT-InSAR) inversion, Gravity Recovery and Climate Experiment/Global Land data Assimilation System (GRACE/GLDAS) modelling, and hydraulic head measurements. The time-series comparison of wells and overall statistics verifies the reliability of both MT-InSAR- and GRACE/GLDAS-derived GWSA, exhibiting a mean R-square of 0.8288 and an RMSE of 7.5 cm. The corrected and integrated GWSA significantly refines the regular grid of ~27 km from GRACE to within 1 km in deep confined aquifers and effectively constrains uncertainties in the GRACE solution, probably caused by spatial leakages in the mass field. Furthermore, the gravity analysis of GWSA in HNP reveals that the central regions in HNP between the Yellow River and the Huaihe River suffer the most severe regional groundwater depletion. The depletion center has shifted to the northwest by 87 km, showing a trend of easing in the north-south direction and expanding in the east-west direction. © 2024, The Authors. All rights reserved.

关键词： Aquifers

来源：评论

学校读者我要写书评

暂无评论

Temporal Convolution Network Based on Attention Mechanism for Well Production Prediction

SSRN

引用

SSRN 2022年

作者： Zhen, Yan Fang, Junyi Zhao, Xiaoming Ge, Jiawang Xiao, Yifei School of Earth Science and Technology Southwest Petroleum University Sichuan Chengdu China Natural Gas Geology Key Laboratory of Sichuan Province Southwest Petroleum University Sichuan Chengdu China Spatial Information Technology and Big Data Mining Research Center Sichuan Chengdu China

Well production forecasting has a very important guiding significance for oilfield production and management. The traditional BP neural network is difficult to deal with the data with time continuity, and the recurrent neural network (RNN) has some shortcomings such as vanishing gradients and exploding gradients. In order to avoid these problems and further improve the accuracy of neural network models in well production prediction. In this study, a prediction model combining temporal convolutional network (TCN) and attention mechanism is proposed. TCN is able to obtain more sensory fields to solve longer time series with fewer network layers by causally expanding convolution. And the residual connectivity in the model structure enables the model to transfer information across layers to solve the problem of network degradation caused by overly deep networks. The attention mechanism can reinforce the influence of important features on the prediction results to improve the prediction accuracy of the model. To evaluate the effectiveness of the model, the TCN-attention model was used to forecast the oil production of two producing wells, and the prediction results were compared with those of LSTM, TCN, RNN and ARIMA. The results showed that TCN-Attention outperformed the other models in the prediction of the two wells in all evaluation metrics. The method proposed in this paper has excellent data fitting ability, can provide accurate prediction results, and have a certain application value in oilfield production. © 2022, The Authors. All rights reserved.

关键词： Time series

来源：评论

学校读者我要写书评

暂无评论

Improved YOLOv8 Algorithm for Small Target Detection in Drone Aerial Photography

引用

Computer Engineering and Applications 2025年第11期61卷 119-131页

作者： Xu, Jingke Suo, Xianglong Zhou, Lei School of Computer Science and Engineering Shenyang Jianzhu University Shenyang110168 China Liaoning Province Big Data Management and Analysis Laboratory of Urban Construction Shenyang110168 China Shenyang Branch of National Special Computer Engineering Technology Research Center Shenyang110168 China

Detecting small and densely distributed targets in UAV aerial images poses challenges such as complex backgrounds, imbalanced sample numbers, and limited computational power. To address these issues, an improved YOLOv8 algorithm, MFF-YOLOv8 (multi-feature fusion YOLOv8), is proposed. This algorithm integrates deformable convolution DCNv3 (deformable convolution v3) into the Bottleneck module of the C2f module to enhance feature extraction. A new MFFPN (multi-feature fusion pyramid network) is designed to add more feature fusion routes, retain low-level feature map details, and improve small target detection. Additionally, a P2 small target detection layer is added and the P5 detection layer is optimized to enhance accuracy and reduce parameters. The introduction of the Dyhead (dynamic head) further improves detection precision of the model. In the experiment on the Visdrone2019 dataset, MFF-YOLOv8s has a 10.2 percentage points and 7.1 percentage points increase in mAP50 and mAP50:95 respectively compared to YOLOv8s, with a 77.04% reduction in parameters. The model's detection accuracy surpasses that of YOLOv11, meeting the precision and lightweight requirements for UAV platforms. © 2025 Journal of Computer Engineering and Applications Beijing Co., Ltd.;science Press. All rights reserved.

关键词： Unmanned aerial vehicles (UAV)

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：