检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Liu, Daizong Yang, Mingyu Qu, Xiaoye Zhou, Pan Cheng, Yu Hu, Wei Wangxuan Institute of Computer Technology Peking University No. 128 Zhongguancun North Street Beijing China The Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan430074 China The Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong

With the significant development of large models in recent years, Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities across a wide range of multimodal understanding and reasoning tasks. Compared to traditional Large Language Models (LLMs), LVLMs present great potential and challenges due to its closer proximity to the multi-resource real-world applications and the complexity of multi-modal processing. However, the vulnerability of LVLMs is relatively underexplored, posing potential security risks in daily usage. In this paper, we provide a comprehensive review of the various forms of existing LVLM attacks. Specifically, we first introduce the background of attacks targeting LVLMs, including the attack preliminary, attack challenges, and attack resources. Then, we systematically review the development of LVLM attack methods, such as adversarial attacks that manipulate model outputs, jailbreak attacks that exploit model vulnerabilities for unauthorized actions, prompt injection attacks that engineer the prompt type and pattern, and data poisoning that affects model training. Finally, we discuss promising research directions in the future. We believe that our survey provides insights into the current landscape of LVLM vulnerabilities, inspiring more researchers to explore and mitigate potential safety issues in LVLM developments. The latest papers on LVLM attacks are continuously collected in https://***/liudaizong/Awesome-LVLM-Attack. Copyright © 2024, The Authors. All rights reserved.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

FedEdge: Accelerating Edge-Assisted Federated Learning 23

FedEdge: Accelerating Edge-Assisted Federated Learning

引用

2023 World Wide Web Conference, WWW 2023

作者： Wang, Kaibin He, Qiang Chen, Feifei Jin, Hai Yang, Yun School of Computer Science and Technology Huazhong University of Science and Technology China Department of Computing Technologies Swinburne University of Technology Australia School of Information Technology Deakin University Australia National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab Huazhong University of Science and Technology Wuhan430074 China

ISBN: (纸本)9781450394161

Federated learning (FL) has been widely acknowledged as a promising solution to training machine learning (ML) model training with privacy preservation. To reduce the traffic overheads incurred by FL systems, edge servers have been included between clients and the parameter server to aggregate clients' local models. Recent studies on this edge-assisted hierarchical FL scheme have focused on ensuring or accelerating model convergence by coping with various factors, e.g., uncertain network conditions, unreliable clients, heterogeneous compute resources, etc. This paper presents our three new discoveries of the edge-assisted hierarchical FL scheme: 1) it wastes significant time during its two-phase training rounds;2) it does not recognize or utilize model diversity when producing a global model;and 3) it is vulnerable to model poisoning attacks. To overcome these drawbacks, we propose FedEdge, a novel edge-assisted hierarchical FL scheme that accelerates model training with asynchronous local federated training and adaptive model aggregation. Extensive experiments are conducted on two widely-used public datasets. The results demonstrate that, compared with state-of-the-art FL schemes, FedEdge accelerates model convergence by 1.14 × -3.20 ×, and improves model accuracy by 2.14% - 6.63%. © 2023 ACM.

关键词： Edge computing

来源：评论

学校读者我要写书评

暂无评论

Discrimination between leucine-rich glioma-inactivated 1 antibody encephalitis and gamma-aminobutyric acid B receptor antibody encephalitis based on ResNet18

引用

Visual Computing for Industry,Biomedicine,and Art 2023年第1期6卷 245-256页

作者： Jian Pan Ruijuan Lv Qun Wang Xiaobin Zhao Jiangang Liu Lin Ai School of Computer and Information Technology Beijing Jiaotong UniversityBeijing 100044China Department of Neurology Beijing Tiantan HospitalCapital Medical UniversityChina National Clinical Research Center for Neurological DiseasesBeijing 100070China Department of Nuclear Medicine Beijing Tiantan HospitalCapital Medical UniversityBeijing 100070China School of Engineering Medicine Beihang UniversityBeijing 100191China Key Laboratory of Big Data-Based Precision Medicine(Beihang University) Ministry of Industry and Information Technology of the People’s Republic of ChinaBeijing 100191China.

This study aims to discriminate between leucine-rich glioma-inactivated 1(LGI1)antibody encephalitis and gammaaminobutyric acid B(GABAB)receptor antibody encephalitis using a convolutional neural network(CNN)model.A total of 81 patients were recruited for this ***18,VGG16,and ResNet50 were trained and tested separately using 3828 positron emission tomography image slices that contained the medial temporal lobe(MTL)or basal ganglia(BG).Leave-one-out cross-validation at the patient level was used to evaluate the CNN *** receiver operating characteristic(ROC)curve and the area under the ROC curve(AUC)were generated to evaluate the CNN *** on the prediction results at slice level,a decision strategy was employed to evaluate the CNN models’performance at patient *** ResNet18 model achieved the best performance at the slice(AUC=0.86,accuracy=80.28%)and patient levels(AUC=0.98,accuracy=96.30%).Specifically,at the slice level,73.28%(1445/1972)of image slices with GABAB receptor antibody encephalitis and 87.72%(1628/1856)of image slices with LGI1 antibody encephalitis were accurately *** the patient level,94.12%(16/17)of patients with GABAB receptor antibody encephalitis and 96.88%(62/64)of patients with LGI1 antibody encephalitis were accurately *** of the image slices extracted using gradient-weighted class activation mapping indicated that the model focused on the MTL and BG for *** general,the ResNet18 model is a potential approach for discriminating between LGI1 and GABAB receptor antibody *** in the MTL and BG is important for discriminating between these two encephalitis subtypes.

关键词： ResNet18 Fluorodeoxyglucose-positron emission tomography GABAB receptor antibody encephalitis Deep learning LGI1 antibody encephalitis

来源：评论

学校读者我要写书评

暂无评论

Out-of-Distribution Fault Diagnosis of Industrial Cyber-Physical Systems Based on Orthogonal Anchor Clustering With Adaptive Balance

IEEE Transactions on Industrial Cyber-Physical Systems

引用

IEEE Transactions on Industrial Cyber-Physical Systems 2025年 3卷 48-60页

作者： Liu, Ruonan Hu, Puyuan Zhao, Siheng Sun, Zhijian Han, Te Pang, Zhibo Zhang, Weidong Shanghai Jiao Tong University Department of Automation Shanghai200240 China National Key Laboratory of Marine Engine Science and Technology Shanghai201108 China Crossocean of Suzhou Technology Suzhou215121 China Beijing Institute of Technology Center for Energy and Environmental Policy Research School of Management Beijing100081 China ABB Corporate Research Sweden Department of Automation Technology Vasteras722 26 Sweden KTH Royal Institute of Technology Department of Intelligent Systems Stockholm114 28 Sweden Henan University Henan Key Laboratory of Big Data Analysis and Processing School of Computer and Information Engineering Kaifeng475001 China Shanghai Jiaotong University Department of Automation Shanghai200240 China

Given the critical role of rotating machinery in industrial cyber-physical systems (ICPS), ensuring their reliable operation is essential for the stability and safety of ICPS. Deep neural networks have demonstrated competitive results for intelligent fault diagnosis, which are usually trained via the historical data of all fault modes. However, in real engineering, it is usually difficult to collect samples and exhaust all failures during the training stage. As a result, out-of-distribution fault diagnosis (OOD-FD) becomes a more realistic problem that requires the methods to not only accurately diagnose the known faults, but also effectively recognize unknown ones. Therefore, a novel orthogonal anchor clustering with focal attention (FA-OAC) is proposed in this paper for OOD-FD. Firstly, an orthogonal anchor clustering (OAC) algorithm is proposed to fix the class center of each fault mode orthogonally and distinguish the known and unknown faults at the class level. Then, because the identifiability of different fault modes changes a lot in OOD-FD, the focal attention mechanism is applied to dynamically adjust the attention to different fault modes according to the distance loss of OAC, thus addressing the identifiability imbalance problem. To verify the effectiveness of the proposed method, different OOD-FD tasks are designed based on two rotating machinery datasets. The experimental results and comparison with state-of-the-art fault diagnosis methods have demonstrated that the proposed method has improved the OOD-FD performance greatly and therefore provides an effective tool in real engineering. © 2023 IEEE.

关键词： Rotating machinery

来源：评论

学校读者我要写书评

暂无评论

Detection and Classification of Transmission Line Transient Faults Based on Graph Convolutional Neural Network

引用

CSEE Journal of Power and Energy Systems 2021年第3期7卷 456-471页

作者： Houjie Tong Robert C.Qiu Dongxia Zhang Haosen Yang Qi Ding Xin Shi Department of Electrical Engineering Center for Big Data and Artificial IntelligenceShanghai Jiao Tong UniversityShanghai 200240China School of Electronic Information and Communication Huazhong University of Science and TechnologyWuhan 430000China Electric Power Research Institute Haidian DistrictBeijing 100192China School of Control and Computer Engineering North China Electric Power UniversityBeijing 102206China

We present a novel transient fault detection and classification approach in power transmission lines based on graph convolutional neural *** with the existing techniques,the proposed approach considers explicit spatial information in sampling sequences as prior knowledge and it has stronger feature extraction *** this basis,a framework for transient fault detection and classification is *** structure is generated to provide topology information to the *** approach takes the adjacency matrix of topology graph and the bus voltage signals during a sampling period after transient faults as inputs,and outputs the predicted classification results ***,the proposed approach is tested in various situations and its generalization ability is verified by experimental *** results show that the proposed approach can detect and classify transient faults more effectively than the existing techniques,and it is practical for online transmission line protection for its rapidness,high robustness and generalization ability.

关键词： Graph convolutional network(GCN) power transmission line fault detection and classification spatio-temporal data topology information

来源：评论

学校读者我要写书评

暂无评论

Reliable Indoor Localization in Multi-Building Environments: Leveraging Environment-Invariant and Position-Related Features

引用

IEEE Internet of Things Journal 2025年

作者： Long, Wenhan Wen, Xinlong Liu, Hao Li, Mengyao Li, Songquan Wu, Yonghui Chen, Fuxiang Liu, Lu Zhu, Rongbo Huazhong Agricultural University College of Informatics Wuhan430070 China Ministry of Education Engineering Research Center of Intelligent Technology for Agriculture Wuhan430070 China Hubei Engineering Technology Research Center of Agricultural Big Data Wuhan430070 China University of Leicester School of Computing and Mathematical Sciences Leicester United Kingdom University of Exeter Department of Computer Science Exeter United Kingdom

Received Signal Strength Indicator (RSSI)-based indoor localization offers a cost-effective solution for autonomous mobile robot navigation in 3D indoor environments, including cross-floor and multi-building structures. However, localization accuracy is fundamentally constrained by the low sampling density and unstable measurement of RSSI data. So far, existing methods neglect cross-environment RSSI coherence (e.g., repeated signal patterns in geometrically similar areas), resulting in unreliable fingerprint databases. What’s more, most approaches fail to model the spatial hierarchy of buildings, floors, and coordinates, which leads to lower accuracy in indoor positioning model predictions. To address these issues, we propose EP-3DLoc, a novel 3D indoor localization framework that combines an Environment-Invariant feature-based data Completion (EIC) method with a Position-Related feature-based Localization (PRL) method. The EIC enhances data quality by filling in sparse RSSI data using environment-invariant features, which are recurring RSSI patterns found in similar environmental structures. The PRL module combines multi-scale RSSI signal processing (raw data and image-like data) with a multi-task network that analyzes location relationships, enhancing localization accuracy in 3D environments. Experimental results on public datasets (TUT2018, UTSIndoorLoc, and UJIIndoorLoc) have demonstrated that EP-3DLoc achieves state-of-the-art performance on indoor localization in multi-building environments. Further testing on the self-constructed dataset HZAUIndoorLoc have revealed that EP-3DLoc not only outperforms existing methods in localization accuracy but also maintains low energy consumption and strong resistance to interference. © 2014 IEEE.

关键词： Floors

来源：评论

学校读者我要写书评

暂无评论

DE-ConvGraph 3D UNet: A Novel Deep Learning Model for Optimizing Radiotherapy Treatment Plans in Oropharyngeal Cancer

DE-ConvGraph 3D UNet: A Novel Deep Learning Model for Optimi...

引用

Applied Imagery Pattern Recognition Workshop (AIPR)

作者： Bhuvanashree Murugadoss J Amudha Vijayan Sugumaran Department of Computer Science and Engineering School of Computing Amrita Vishwa Vidyapeetham Bengaluru Karnataka India Center for Data Science and Big Data Analytics Oakland University Rochester MI USA Department of Decision and Information Sciences Oakland University Rochester MI USA

Automated radiotherapy treatment planning aims to improve treatment accuracy and efficiency. However, the prevalent Knowledge-Based Planning (KBP) method faces issues like lengthy manual problem formulation and challenges in accurately modeling human anatomy and processing high-dimensional data. This work focuses on treatment plan optimization and considers deep learning as a potential solution. Deep learning algorithms, by virtue of their ability to learn from large quantities of data and model complex relationships, can automate the formulation of the optimization problem in KBP, saving significant time and effort. However, despite these compelling advantages, it should be noted that the current convolution-based encoder-decoder models used for radiotherapy treatment plan optimization have a limited capability in capturing long-range dependencies between distant voxels. This work aims to introduce ‘DE-ConvGraph 3D UNet’, a novel deep learning model, to address these limitations and optimize radiotherapy treatment plans for oropharyngeal cancer. The proposed ‘DE-ConvGraph 3D UNet’ model includes a Graph Convolutional Network (GCN) component to capture long-range dependencies between distant voxels. Furthermore, the dual-encoder structure of the model combines the strengths of GCN and 3D Convolutional U-Net, enabling global relationships and local pattern capturing in a 3D patient volume. Experiments were conducted using the benchmark dataset - OpenKBP-Opt. The proposed model shows improvements in performance in comparison to state-of-the-art U-Net variants in terms of mean squared voxel-wise error, dose volume histogram points and clinical criteria satisfaction.

关键词：

来源：评论

学校读者我要写书评

暂无评论

How graph neural networks learn: lessons from training dynamics 24

How graph neural networks learn: lessons from training dynam...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Chenxiao Yang Qitian Wu David Wipf Ruoyu Sun Junchi Yan School of Artificial Intelligence & Department of Computer Science and Engineering & MoE Lab of AI Shanghai Jiao Tong University Amazon Web Services School of Data Science The Chinese University of Hong Kong Shenzhen and Shenzhen International Center for Industrial and Applied Mathematics Shenzhen Research Institute of Big Data

A long-standing goal in deep learning has been to characterize the learning behavior of black-box models in a more interpretable manner. For graph neural networks (GNNs), considerable advances have been made in formalizing what functions they can represent, but whether GNNs will learn desired functions during the optimization process remains less clear. To fill this gap, we study their training dynamics in function space. In particular, we find that the gradient descent optimization of GNNs implicitly leverages the graph structure to update the learned function, as can be quantified by a phenomenon which we call kernel-graph alignment. We provide theoretical explanations for the emergence of this phenomenon in the overparameterized regime and empirically validate it on real-world GNNs. This finding offers new interpretable insights into when and why the learned GNN functions generalize, highlighting their limitations in heterophilic graphs. Practically, we propose a parameter-free algorithm that directly uses a sparse matrix (i.e. graph adjacency) to update the learned function. We demonstrate that this embarrassingly simple approach can be as effective as GNNs while being orders-of-magnitude faster.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Snapshot boosting: a fast ensemble framework for deep neural networks

引用

Science China(Information Sciences) 2020年第1期63卷 77-88页

作者： Wentao ZHANG Jiawei JIANG Yingxia SHAO Bin CUI Center for Data Science Peking University National Engineering Laboratory for Big Data Analysis and Applications Department of Computer Science Beijing Key Lab of Intelligent Telecommunications Software and Multimedia School of Computer ScienceBeijing University of Posts and Telecommunications Key Lab of High Confidence Software Technologies Department of Computer SciencePeking University

Boosting has been proven to be effective in improving the generalization of machine learning models in many fields. It is capable of getting high-diversity base learners and getting an accurate ensemble model by combining a sufficient number of weak learners. However, it is rarely used in deep learning due to the high training budget of the neural network. Another method named snapshot ensemble can significantly reduce the training budget, but it is hard to balance the tradeoff between training costs and diversity. Inspired by the ideas of snapshot ensemble and boosting, we propose a method named snapshot boosting. A series of operations are performed to get many base models with high diversity and accuracy, such as the use of the validation set, the boosting-based training framework, and the effective ensemble strategy. Last, we evaluate our method on the computer vision(CV) and the natural language processing(NLP) tasks, and the results show that snapshot boosting can get a more balanced trade-off between training expenses and ensemble accuracy than other well-known ensemble methods.

关键词： ensemble learning deep learning boosting neural network snapshot ensemble

来源：评论

学校读者我要写书评

暂无评论

One-Dimensional EEG Artifact Removal Network Based on Convolutional Neural Networks

Journal of Network Intelligence

引用

Journal of Network Intelligence 2024年第1期9卷 142-159页

作者： Xiong, Jun Meng, Xiang-Long Chen, Zhao-Qi Wang, Chuan-Sheng Zhang, Fu-Quan Grau, Antoni Chen, Yang Huang, Jing-Wei School of Computer and Data Science Minjiang University Fuzhou University Town No. 200 Xiyuangong Road Fuzhou China College of Electronic Engineering Shandong University of Science and Technology No. 579 Qianwangang Road Huangdao District Qingdao China College of Computer and Big Data Fuzhou University Fuzhou University Town No. 2 Wulongjiang North Road Fuzhou China Department of Automatic Control Technical Polytechnic University of Catalonia Autonomous Region of Catalonia Barcelona Spain Digital Media Art Key Laboratory of Sichuan Province Sichuan Conservatory of Music Fuzhou Technology Innovation Center of intelligent Manufacturing information System Minjiang University Fuzhou University Town No. 200 Xiyuangong Road Fuzhou China Fujian Province University No. 1 Campus New Village Longjiang Street Fuqing China School of Mechanical and Automotive Engineering Fujian University of Technology No. 33 Xuefu South Road University New District Fuzhou China

The electroencephalogram (EEG) serves as a significant tool in the realms of clinical medicine, cerebral investigation, and neurological disorders research. However, the EEG records we obtain are often easily contaminated by various artifacts, which can blur or distort the underlying EEG signals and make data interpretation difficult. Generally speaking, removing EEG artifacts is considered an essential step in brain signal analysis. Therefore, removing artifacts is crucial for obtaining accurate and reliable EEG signals for subsequent analysis. Recently, deep learning techniques have found widespread application across various domains for denoising tasks, including image denoising and EEG denoising. Many advanced algorithms have been developed in image denoising, which has achieved good results in enhancing low-quality images. Moreover, it has shown superior performance in EEG denoising. In contrast, few people have devoted themselves to studying EEG denoising, and existing convolutional neural network EEG denoising methods still have problems of overfitting and poor denoising effect in Electromyograph(EMG) and ElectroOculoGram(EOG) artifact removal. Therefore, this paper proposes a method called DWINet (De-artifacting with Image-based Network for EEG Signals) based on an image dehazing network DRHNet for removing artifacts from EEG signals. Specifically, our approach DWINet, addresses the de-artifacting issue in EEG signals by converting it as an image dehazing problem and utilizes the image dehazing capability of DRHNet to enhance the denoising performance of EEG signals. Experimental results demonstrate that the proposed method outperforms the compared algorithms in removing the ocular artifact in EEG signals and exhibits higher accuracy and robustness. © 2024, J. Network Intell. All rights reserved.

关键词： Image enhancement

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：