检索结果-内蒙古大学图书馆

4th International Conference on Machine Learning for Cyber security, ML4CS 2022

作者： Lin, Jianchen Yao, Yanqing State Key Laboratory of Software Development Environment Beihang University Beijing100191 China State Key Laboratory of Cryptology Beijing100878 China Key Laboratory of Aerospace Network Security Ministry of Industry and Information Technology School of Cyber Science and Technology Beihang University Beijing100191 China State Grid Fujian Fuzhou350100 China

ISBN: (纸本)9783031200953

As a branch of deep learning, conditional Generative Adversarial Net (cGAN) has achieved remarkable results in target-oriented image generation, but the process of training the model lacks privacy guarantee. Differential Privacy (DP) is a privacy-preserving mechanism that could be utilized in the Stochastic Gradient Descent (SGD) of deep learning to maintain the confidentiality of training data. However, the quality and diversity of differentially private conditional image synthesis remain large room for improvement because traditional mechanisms with thick granularities and rigid clipping bounds in Differentially Private SGD (DPSGD) could lead to huge performance loss. To tackle this, we propose three algorithms with different clipping granularities and clipping bound distributions, and investigate the effects on enhancing conditional image generation under Differentially Private Projection cGAN (DP-PcGAN). We compare and analyze the performance of our strategies with methods concerned in terms of seven metrics, finding ratio-based gradient clipping highly effective and robust in training DP-PcGAN for generating visually promising images. Specifically, the Per-layer RGC surpasses previous methods under all evaluation metrics utilizing the second least privacy expense for MNIST dataset, while Per-unit RGC attains encouraging outcomes with a moderate converging speed under FashionMNIST dataset. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

Improved Density-Based data Association Method Under Azimuth Rate Constraint

Improved Density-Based Data Association Method Under Azimuth...

引用

2024 OES China Ocean Acoustics, COA 2024

作者： Qi, Bin Hui, Chenxin Hao, Yu Li, Xiang Mathiopoulos, P. Takis Harbin Engineering University National Key Laboratory of Underwater Acoustic Technology Harbin150001 China Ministry of Industry and Information Technology Key Laboratory of Marine Information Acquisition and Security Harbin150001 China College of Underwater Acoustic Engineering Harbin Engineering University Harbin150001 China National and Kapodistrian University of Athens Department of Informatics and Telecommunications Greece

ISBN: (纸本)9798350340297

Passive sonar has been extensively researched in recent years. However, bearing-only tracking faces significant challenges because passive detection only provides target information. Traditional algorithms estimate target states using Rao-Blackwellized Monte Carlo data Association and form target trajectories using clustering and label management algorithms. However, the tracking effectiveness of this method depends on the results of the clustering algorithm, which leads to trajectory discontinuity in scenarios with target trajectory crossings. This study addresses the problem of trajectory crossing in passive scenarios by proposing an improved clustering method and label management logic. The experimental results confirmed that the proposed algorithm in this scenario can accurately estimate the number of targets and achieve robust target tracking. © 2024 IEEE.

关键词： Trajectories

来源：评论

学校读者我要写书评

暂无评论

Cross-modality information Check for Detecting Jailbreaking in Multimodal Large Language Models

arXiv

引用

arXiv 2024年

作者： Xu, Yue Qi, Xiuyuan Qin, Zhan Wang, Wenjie School of Information Science and Technology ShanghaiTech University China The State Key Laboratory of Blockchain and Data Security Zhejiang University China

Multimodal Large Language Models (MLLMs) extend the capacity of LLMs to understand multimodal information comprehensively, achieving remarkable performance in many vision-centric tasks. Despite that, recent studies have shown that these models are susceptible to jailbreak attacks, which refer to an exploitative technique where malicious users can break the safety alignment of the target model and generate misleading and harmful answers. This potential threat is caused by both the inherent vulnerabilities of LLM and the larger attack scope introduced by vision input. To enhance the security of MLLMs against jailbreak attacks, researchers have developed various defense techniques. However, these methods either require modifications to the model’s internal structure or demand significant computational resources during the inference phase. Multimodal information is a double-edged sword. While it increases the risk of attacks, it also provides additional data that can enhance safeguards. Inspired by this, we propose Cross-modality information DEtectoR (CIDER), a plug-and-play jailbreaking detector designed to identify maliciously perturbed image inputs, utilizing the cross-modal similarity between harmful queries and adversarial images. CIDER is independent of the target MLLMs and requires less computation cost. Extensive experimental results demonstrate the effectiveness and efficiency of CIDER, as well as its transferability to both white-box and black-box MLLMs. The resource is available at https://***/PandragonXIII/CIDER. Copyright © 2024, The Authors. All rights reserved.

关键词： Wine

来源：评论

学校读者我要写书评

暂无评论

Semisupervised Encrypted Traffic Identification Based on Auxiliary Classification Generative Adversarial Network

引用

Computer Systems Science & Engineering 2021年第12期39卷 373-390页

作者： Jiaming Mao Mingming Zhang Mu Chen Lu Chen Fei Xia Lei Fan ZiXuan Wang Wenbing Zhao State Grid Jiangsu Electric Power Co. Ltd.Information and Telecommunication BranchNanJing210003China Institute of Information and Communication Global Energy Interconnection Research InstituteState Grid Key Laboratory of Information and Network SecurityNanjing210003China Engineering Research Center of Post Big Data Technology and Application of Jiangsu Province Research and Development Center of Post Industry Technology of the State Posts Bureau(Internet of Things Technology)Broadband Wireless Communication Technology Engineering Research Center of the Ministry of EducationNanjing University of Posts and TelecommunicationsNanjing210003China Department of Electrical Engineering and Computer Science Cleveland State UniversityCleveland44115USA

The rapidly increasing popularity of mobile devices has changed the methods with which people access various network services and increased net-work traffic *** the past few decades,network traffic identification has been a research hotspot in the field of network management and security ***,as more network services use encryption technology,network traffic identification faces many *** classic machine learning methods can solve many problems that cannot be solved by port-and payload-based methods,manually extract features that are frequently updated is time-consuming and *** learning has good automatic feature learning capabilities and is an ideal method for network traffic identification,particularly encrypted traffic identification;Existing recognition methods based on deep learning primarily use supervised learning methods and rely on many labeled ***,in real scenarios,labeled samples are often difficult to *** paper adjusts the structure of the auxiliary classification generation adversarial network(ACGAN)so that it can use unlabeled samples for training,and use the wasserstein distance instead of the original cross entropy as the loss function to achieve semisupervised *** results show that the identification accuracy of ISCX and USTC data sets using the proposed method yields markedly better performance when the number of labeled samples is small compared to that of convolutional neural network(CNN)based classifier.

关键词： Encrypted traffic recognition deep learning generative adversarial network traffic classification semisupervised learning

来源：评论

学校读者我要写书评

暂无评论

Multi-domain feature analysis of underwater acoustic communication signals with different modulation techniques 3

Multi-domain feature analysis of underwater acoustic communi...

引用

3rd International Conference on Signal Image Processing and Communication, ICSIPC 2023

作者： Wang, Yang Wang, Tao Shen, Tongsheng Qiao, Gang Zhou, Feng National Key Laboratory of Underwater Acoustic Technology Harbin Engineering University Harbin150001 China Key Laboratory of Marine Information Acquisition and Security Harbin Engineering University Ministry of Industry and Information Technology Harbin150001 China College of Underwater Acoustic Engineering Harbin Engineering University Harbin150001 China National Innovation Institute of Defense Technology Beijing100071 China

ISBN: (纸本)9781510670945

Strongly robust signal feature construction is paramount to improve the accuracy of underwater acoustic (UWA) communication signal modulation recognition. But the ambient noise in the underwater channel confuses the signal features, and the single feature extraction technique poses huge difficulties to the modulation identification. In this paper, various feature extraction methods are advanced for several commonly used modulation signals in UWA communication systems. We analyze the differences in their features, and validate the advantages of multi-feature fusion in characterizing signal differences. Firstly, the time-frequency maps of different signals are gained by continuous wavelet transform (CWT) theory. Secondly, we propose the two strongest spectral lines extraction method based on power spectrum and autocorrelation spectrum, and the second strongest spectral line amplitude is used as a new and more obvious feature parameter. Thirdly, two correlation coefficient features based on circular spectral are constructed. Through analyzing the simulation data, the results show that different signal feature extraction techniques provide richer analysis domains and the multi-domain fusion features have more robust performance under signal-to-noise ratio (SNR) of 3dB. © 2023 SPIE.

关键词： Underwater acoustics

来源：评论

学校读者我要写书评

暂无评论

Target detection algorithm in side-scan sonar image based on pixel importance value measurement

引用

Chinese Journal of Acoustics 2020年第3期39卷 329-340页

作者： BIAN Hongyu CHEN Yiming ZHANG Zhigang WEI Mingzhe Acoustic Science and Technology Laboratory Harbin Engineering UniversityHarbin 150001 Peng Cheng Laboratory Shenzhen 518000 Key Laboratory of Marine Information Acquisition and Security(Harbin Engineering University) Ministry of Industry and Information TechnologyHarbin 150001 College of Underwater Acoustic Engineering Harbin Engineering UniversityHarbin 150001 Intelligence and Information Engineering College Tangshan UniversityTangshan 063000

Side-scan sonar detection application always combines with unstable results.A two-stage novel pixel importance value measurement algorithm is proposed to stabilize the detection ability and false alarm probability *** first stage of the algorithm,a new feature defined as pixel importance value(PIV)is proposed in terms of distances between the target pixel and each other *** measurement of current pixel is defined as the weighted sum of all remaining segmented *** weighted part refers to Gaussian kernel,which means closer pixels gets higher ***,targets with higher PIV can be *** the second stage,we use convolutional neural network as classifier to eliminate the dot-like false *** experiment data is obtained by autonomous underwater vehicle,where we demonstrate superior performance of our algorithm over the state-of-the-art sonar detection algorithms in terms of 90.39% recall rate and 2.39% false alarm probability.

关键词： algorithm false value

来源：评论

学校读者我要写书评

暂无评论

Suppression of Cross-Terms in the Wigner-Ville Distribution of Underwater Static Target Echo

Suppression of Cross-Terms in the Wigner-Ville Distribution ...

引用

2024 OES China Ocean Acoustics, COA 2024

作者： Xu, Wencheng Li, Xiukun Yu, Ge Harbin Engineering University National Key Laboratory of Underwater Acoustic Technology China Harbin Engineering University Ministry of Industry and Information Technology Key Laboratory of Marine Information Acquisition and Security China Harbin Engineering University College of Underwater Acoustic Engineering Harbin China Engineering University Sanya Nanhai Innovation and Development Base of Harbin Sanya Harbin China

ISBN: (纸本)9798350340297

The echo information components of underwater static targets are intricate and encompass vital details regarding target shape and structure. When the active sonar emission signal adopts a linear frequency modulation (LFM) signal, the target echo comprises superimpositions of multiple LFM signals with varying amplitudes and delays. Although the Wigner-Ville distribution (WVD) serves as a potent tool for analyzing LFM signals, the WVD for multicomponent LFM signals yields numerous cross-terms. These cross-terms impede further assessment of signal quantity and information content. Consequently, in this study, a novel approach to suppress cross-terms is proposed. By extracting the WVD along the frequency slope direction, the extracted results undergo filtering via a window function and are subsequently multiplied by a summation suppression factor to further attenuate cross-terms and noise. Moreover, in this study, a comparative analysis of different window functions is conducted. Simulation experiments and practical data processing confirm the effectiveness of the proposed method in this study. © 2024 IEEE.

关键词： Wigner-Ville distribution

来源：评论

学校读者我要写书评

暂无评论

Navigation Calibration Algorithm Assisted by Single Beacon

Navigation Calibration Algorithm Assisted by Single Beacon

引用

IEEE International Conference on Signal Processing, Communications and Computing

作者： Jucheng Zhang Lan Qin Dajun Sun Yunfeng Han Acoustic Science and Technology Laboratory (Harbin Engineering University) Key Laboratory of Marine Information Acquisition and Security (Ministry of Industry and Information Technology) Harbin China

Autonomous underwater vehicle has become an important tool to replace human beings to complete abyssal operations and missions. As the most widely used positioning and navigation technology, inertial navigation system integrated with doppler velocity log to carry out dead reckoning exists significant shortcomings of positioning error accumulated over time. In this paper, on the background of the modification and calibration of inertial navigation drift, a positioning and navigation algorithm based on the principle of indirect adjustment is proposed, which takes single-beacon range information and velocity information as the observation quantity. This paper builds a single-beacon navigation model and analyzes its observability, as well as evaluates the positioning accuracy. The simulation analysis and lake test data processing show that the algorithm can effectively suppress the divergence of inertial navigation error and realize the AUV's long-time high-precision navigation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Design and Verification of Wake-up Signal for Underwater Nodes

Design and Verification of Wake-up Signal for Underwater Nod...

引用

IEEE International Conference on Signal Processing, Communications and Computing

作者： Dajun Sun YujieOuyang Yunfeng Han Jucheng Zhang Acoustic Science and Technology Laboratory (Harbin Engineering University) Key Laboratory of Marine Information Acquisition and Security (Ministry of Industry and Information Technology) Harbin China

The construction and improvement of the underwater acoustic network is the premise and guarantee for the development of the marine industry. Because the underwater nodes need to work for a long time, it is especially important to ensure that the nodes have a long standby capacity. In general, the node is in a low-power standby state waiting for a wake-up signal. When the node detects the wake-up signal, it will resume normal operation. In this paper, we propose a signal design based on the m-sequence. which can detect the hidden awakening signal in the complex environment with low SNR and small Doppler shift. Simulation and experimental data indicate that when the input SNR is as low as -11dB and the signal has a small Doppler shift, the system can still achieve a detection probability of 100% and ensure that the false alarm probability is lower than 10~(-6).

关键词：

来源：评论

学校读者我要写书评

暂无评论

Disentangling the Individual-Shared and Individual-Specific Subspace of Altered Brain Functional Connectivity in Autism Spectrum Disorder (Vol 95, pg 870, 2024)

引用

BIOLOGICAL PSYCHIATRY 2024年第9期95卷 909-909页

作者： Shan [a]Sichuan Provincial Center for Mental Health Sichuan Provincial People's Hospital School of Life Science and Technology University of Electronic Science and Technology of China Chengdu China [b]Ministry of Education Key Lab for Neuroinformation High-Field Magnetic Resonance Brain Imaging Key Laboratory of Sichuan Province University of Electronic Science and Technology of China Chengdu China [c]Department of Psychiatry and Biobehavioral Sciences University of California Los Angeles Los Angeles California [d]College of Blockchain Industry Chengdu University of Information Technology Chengdu China

Background Despite considerable effort toward understanding the neural basis of autism spectrum disorder (ASD) using case-control analyses of resting-state functional magnetic resonance imaging data, findings are often not reproducible, largely due to biological and clinical heterogeneity among individuals with ASD. Thus, exploring the individual-shared and individual-specific altered functional connectivity (AFC) in ASD is important to understand this complex, heterogeneous disorder. Methods We considered 254 individuals with ASD and 295 typically developing individuals from the Autism Brain Imaging data Exchange to explore the individual-shared and individual-specific subspaces of AFC. First, we computed AFC matrices of individuals with ASD compared with typically developing individuals. Then, common orthogonal basis extraction was used to project AFC of ASD onto 2 subspaces: an individual-shared subspace, which represents altered connectivity patterns shared across ASD, and an individual-specific subspace, which represents the remaining individual characteristics after eliminating the individual-shared altered connectivity patterns. Results Analysis yielded 3 common components spanning the individual-shared subspace. Common components were associated with differences of functional connectivity at the group level. AFC in the individual-specific subspace improved the prediction of clinical symptoms . The default mode network–related and cingulo-opercular network–related magnitudes of AFC in the individual-specific subspace were significantly correlated with symptom severity in social communication deficits and restricted, repetitive behaviors in ASD. Conclusions Our study decomposed AFC of ASD into individual-shared and individual-specific subspaces, highlighting the importance of capturing and capitalizing on individual-specific brain connectivity features for dissecting heterogeneity. Our analysis framework provides a blueprint for parsing heterogeneity in other

关键词： Altered functional connectivity ASD heterogeneity Common orthogonal basis 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：