检索结果-内蒙古大学图书馆

41st International Conference on Machine Learning, ICML 2024

作者： Lin, Runqi Yu, Chaojian Han, Bo Su, Hang Liu, Tongliang Sydeny AI Centre School of Computer Science The University of Sydney Sydney Australia Department of Computer Science Hong Kong Baptist University Hong Kong Department of Computer Science and Technology Institute for AI BNRist Center Tsinghua University Beijing China

Catastrophic overfitting (CO) presents a significant challenge in single-step adversarial training (AT), manifesting as highly distorted deep neural networks (DNNs) that are vulnerable to multi-step adversarial attacks. However, the underlying factors that lead to the distortion of decision boundaries remain unclear. In this work, we delve into the specific changes within different DNN layers and discover that during CO, the former layers are more susceptible, experiencing earlier and greater distortion, while the latter layers show relative insensitivity. Our analysis further reveals that this increased sensitivity in former layers stems from the formation of pseudo-robust shortcuts, which alone can impeccably defend against single-step adversarial attacks but bypass genuine-robust learning, resulting in distorted decision boundaries. Eliminating these shortcuts can partially restore robustness in DNNs from the CO state, thereby verifying that dependence on them triggers the occurrence of CO. This understanding motivates us to implement adaptive weight perturbations across different layers to hinder the generation of pseudo-robust shortcuts, consequently mitigating CO. Extensive experiments demonstrate that our proposed method, Layer-Aware Adversarial Weight Perturbation (LAP), can effectively prevent CO and further enhance robustness. Our implementation can be found at https://***/tmllab/2024_ICML_LAP. Copyright 2024 by the author(s)

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

A Survey of Cyber Range: Current Status, Analysis, and Future Trends 2nd

A Survey of Cyber Range: Current Status, Analysis, and Futur...

引用

2nd International Conference on network Simulation and Evaluation, NSE 2023

作者： Zhang, Jiawei Li, Zongzhe Chen, Rui Han, Weihong Hu, Ning Zhang, Yanchun Department of New Network Peng Cheng Laboratory Shenzhen China School of Computer Science and Technology Zhejiang Normal University Jinhua China

ISBN: (纸本)9789819745210

The escalating threats in the realm of cybersecurity have given rise to the demand for security assessments of information systems. However, due to the limitations of freely and continuously conducting real-time security tests on actual systems, the need for simulation-based evaluation arises, leading to the emergence of network cyber range technology. In recent years, network cyber ranges have evolved into pivotal scientific infrastructures supporting the construction of national cybersecurity strategies. They are rapidly becoming the digital innovation frontier for countries and organizations such as the United States, the European Union, and Japan. These cyber ranges effectively support a range of activities in the field of cybersecurity, including security technology research, red team-blue team exercises, security capability evaluations, and validation of emerging network technologies. In this paper, we collect, consolidate, and summarize existing cyber ranges, analyze their core technologies, key functionalities, and classical applications. Furthermore, we provide an overview of the current state of cyber range technology and shed light on its shortcomings and future development trends. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Cybersecurity

来源：评论

学校读者我要写书评

暂无评论

Research on Classroom Behavior Target Detection Algorithm based on YOLOv7 6

Research on Classroom Behavior Target Detection Algorithm ba...

引用

6th IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference, IMCEC 2024

作者： Qiu, Wenke Wu, Xiaoming Liu, Xiangzhi Computer Science and Qilu University of Technology Shandong Academy of Sciences ShanDong Jinan China Qilu University of Technology Shandong Academy of Sciences Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center National Supercomputer Center in Jinan Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computer Networks Jinan China

ISBN: (纸本)9798350316520

The integration of video information for the purpose of detecting students' behavior in the classroom, with the aim of enhancing teaching efficiency and maintaining a stable classroom environment, offers significant assistance. However, existing models usually struggle to achieve good accuracy due to the lack of datasets, dense targets, interclass occlusion, and other issues. To address these difficulties, the CD-YOLO classroom behavior detection algorithm is proposed in this paper. And the classroom behavior dataset (DSL-Dataset) is created. We propose a loss function named VAIoU, which can be combined with vector ratio to filter the prediction box. To further improve the detection accuracy of occluded targets, we propose an attention module CM module with command feature fusion. Moreover, in this paper, we improve the convolutional kernel to design a lightweight convolutional kernel FFconv. Experimental results on DSL-Dataset show that CD-YOLO outperforms existing one-stage detection models in terms of detection accuracy. © 2024 IEEE.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

On a novel tracking differentiator design based on iterative learning in a moving window

引用

Control Theory and technology 2023年第1期21卷 46-55页

作者： Xiangyang Li Rafal Madonski Zhiqiang Gao Senping Tian Key Laboratory of Autonomous Systems and Network Control Ministry of EducationSchool of Automation Science and EngineeringSouth China University of TechnologyGuangzhou 510641China Energy and Electricity Research Center Jinan UniversityZhuhai 519070China Department of Electrical and Computer Science Cleveland State UniversityClevelandOH 44115USA

Differential signals are key in control engineering as they anticipate future behavior of process variables and therefore are critical in formulating control laws such as proportional-integral-derivative(PID).The practical challenge,however,is to extract such signals from noisy measurements and this difficulty is addressed first by *** in the form of linear and nonlinear tracking differentiator(TD).While improvements were made,TD did not completely resolve the conflict between the noise sensitivity and the accuracy and timeliness of the *** two approaches proposed in this paper start with the basic linear TD,but apply iterative learning mechanism to the historical data in a moving window(MW),to form two new iterative learning tracking differentiators(IL-TD):one is a parallel IL-TD using an iterative ladder network structure which is implementable in analog circuits;the other a serial IL-TD which is implementable digitally on any computer *** algorithms are validated in simulations which show that the proposed two IL-TDs have better tracking differentiation and de-noise performance compared to the existing linear TD.

关键词： Tracking differentiator(TD) Iterative learning Iterative learning tracking differentiator(IL-TD) Active disturbance rejection control(ADRC)-Two-dimensional system(2-D system)

来源：评论

学校读者我要写书评

暂无评论

Enhancing Skin Cancer Diagnosis with Deep Learning:A Hybrid CNN-RNN Approach

引用

computers, Materials & Continua 2024年第4期79卷 1497-1519页

作者： Syeda Shamaila Zareen Guangmin Sun Mahwish Kundi Syed Furqan Qadri Salman Qadri Faculty of Information Technology Beijing University of TechnologyBeijing100124China Computer Science International Engineering Collage Maynooth UniversityKildareW23 F2H6Irland Research Center for Data Hub and Security Zhejiang LabHangzhou311121China Computer Science Department MNS University of AgricultureMultan59220Pakistan

Skin cancer diagnosis is difficult due to lesion presentation variability. Conventionalmethods struggle to manuallyextract features and capture lesions spatial and temporal variations. This study introduces a deep learning-basedConvolutional and Recurrent Neural network (CNN-RNN) model with a ResNet-50 architecture which usedas the feature extractor to enhance skin cancer classification. Leveraging synergistic spatial feature extractionand temporal sequence learning, the model demonstrates robust performance on a dataset of 9000 skin lesionphotos from nine cancer types. Using pre-trained ResNet-50 for spatial data extraction and Long Short-TermMemory (LSTM) for temporal dependencies, the model achieves a high average recognition accuracy, surpassingprevious methods. The comprehensive evaluation, including accuracy, precision, recall, and F1-score, underscoresthe model’s competence in categorizing skin cancer types. This research contributes a sophisticated model andvaluable guidance for deep learning-based diagnostics, also this model excels in overcoming spatial and temporalcomplexities, offering a sophisticated solution for dermatological diagnostics research.

关键词： Skin cancer classification deep learning Convolutional Neural network(CNN) RNN ResNet-50

来源：评论

学校读者我要写书评

暂无评论

Inverse Scattering of Two-dimensional Circular Cylindrical Conductors by Convolutional Neural network 4

Inverse Scattering of Two-dimensional Circular Cylindrical C...

引用

4th IEEE International Conference on Electronic Communications, Internet of Things and Big Data, ICEIB 2024

作者： Chen, Po-Hsiang Hsu, Hsiu-Hui Chiu, Chien-Ching Chien, Wei Yu, Kai-Xu Tamkang University Department of Electrical and Computer Engineering New Taipei City Taiwan Lunghwa University of Science and Technology Department of Computer Information and Network Engineering Taoyuan City Taiwan

ISBN: (纸本)9798350360721

We used a Convolutional Neural network (CNN) to reconstruct the electromagnetic imaging of a perfect conductor in half-space through the traverse magnetic wave. Since the conventional iterative algorithm is time-consuming, we developed a real-time electromagnetic imaging technique to apply to the CNN to reconstruct the half-space conducting cylinders. The sizes and positions of the cylinders were calculated through the Direct Sampling Method (DSM) in the scattered field. The original ground truth was fed to CNN for training to predict accurate sizes and positions of the shape. The accurate size and position of the shape for the perfect conductor were reconstructed. © 2024 IEEE.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

OUT-OF-DOMAIN UNLABELED DATA IMPROVES GENERALIZATION 12

OUT-OF-DOMAIN UNLABELED DATA IMPROVES GENERALIZATION

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Saberi, Amir Hossein Movasaghinia, Mohammad Hosein Najafi, Amir Motahari, Seyed Abolfazl Heidari, Alireza Khalaj, Babak H. Department of Electrical Engineering Department of Computer Engineering Sharif Center for Information Systems and Data Science Iran Sharif Institute for Convergence Science & Technology Sharif University of Technology Tehran Iran

We propose a novel framework for incorporating unlabeled data into semi-supervised classification problems, where scenarios involving the minimization of either i) adversarially robust or ii) non-robust loss functions have been considered. Notably, we allow the unlabeled samples to deviate slightly (in total variation sense) from the in-domain distribution. The core idea behind our framework is to combine Distributionally Robust Optimization (DRO) with self-supervised training. As a result, we also leverage efficient polynomial-time algorithms for the training stage. From a theoretical standpoint, we apply our framework on the classification problem of a mixture of two Gaussians in Rd, where in addition to the m independent and labeled samples from the true distribution, a set of n (usually with n m) out of domain and unlabeled samples are given as well. Using only the labeled data, it is known that the generalization error can be bounded by ∝ (d/m)1/2. However, using our method on both isotropic and non-isotropic Gaussian mixture models, one can derive a new set of analytically explicit and non-asymptotic bounds which show substantial improvement on the generalization error compared to ERM. Our results underscore two significant insights: 1) out-of-domain samples, even when unlabeled, can be harnessed to narrow the generalization gap, provided that the true data distribution adheres to a form of the "cluster assumption", and 2) the semi-supervised learning paradigm can be regarded as a special case of our framework when there are no distributional shifts. We validate our claims through experiments conducted on a variety of synthetic and real-world datasets. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Polynomial approximation

来源：评论

学校读者我要写书评

暂无评论

A Novel Congestion Control Scheme Using Fuzzy Logic Systems to Enhance the Path Selection Criteria in Routing Protocols for Low-Power and Lossy networks on the Internet of Things

引用

SN computer science 2024年第5期5卷 1-31页

作者： Aimtongkham, Phet Musikawan, Pakarat Kongsorot, Yanika So-In, Chakchai Applied Network Technology (ANT) Department of Computer Science College of Computing Khon Kaen University Khon Kaen 40002 Thailand Advanced Intelligent Interdisciplinary Integration (AIII) Department of Computer Science College of Computing Khon Kaen University Khon Kaen 40002 Thailand

In routing protocols for low-power and lossy (RPL)-based Internet of Things (IoT) networks, congestion control is essential for ensuring efficient and dependable communications with energy awareness. As the number of devices and the amount of data traffic increase, congestion typically occurs, resulting in degraded network performance, i.e., increased packet losses and delays with decreased energy efficiency. Many parameters relating to congestion behaviors have been considered and dynamically changed, as have the possible transmission paths either within a single destination-oriented directed acyclic graph (DoDAG) or across different DoDAGs. Thus, this research investigates congestion mitigation methods for RPL-IoT networks based on a fuzzy logic system (FLS) by proposing a novel FLS scheme to determine the selection criteria for conducting path management in three FLS components: (1) inside/outside the DoDAG selection process, (2) inside the path selection process within a single DoDAG (the so-called inside DoDAG) for high or to-be-high loads, and (3) inside the path selection process across the DoDAGs (the so-called outside DoDAG) during severe congestion applying the concept of a multi-DoDAG node (MDN) to determine both a suitable MDN and data traffic for alternate path selection. Simulation results demonstrate that the proposed method effectively controls congestion levels, decreases packet losses and latency, and attains improved energy efficiency in comparison with the state-of-the-art congestion control schemes in RPL-IoT networks. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Congestion controls Fuzzy logic systems Internet of Things Objective function Path selection Routing protocol for low-power and lossy networks

来源：评论

学校读者我要写书评

暂无评论

Robust Voice Activity Detection using Locality-Sensitive Hashing and Residual Frequency-Temporal Attention 25

Robust Voice Activity Detection using Locality-Sensitive Has...

引用

25th Interspeech Conferece 2024

作者： Li, Shu Zhang, Peng Li, Ye Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center National Supercomputer Center in Jinan Qilu University of Technology Shandong Academy of Sciences China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science China

For voice activity detection (VAD), recent works focus on learning the attention distribution on contextual information of speech to reduce the impact of irrelevant noise. However, contextual frames selected with specific steps may not be relevant, and these attention mechanisms can not fully discover the structure and characteristics of speech. In this paper, we explore a self-attention-inspired locality-sensitive hashing algorithm (SALSH) for dynamic and efficient contextual frame selection to enrich the frame-level features into a 2D partial spectrogram. Then, we propose a residual frequency-temporal attention model (FTAM) for VAD, consisting of an interval branch, an analogous hourglass structure with channel attention, and an attention learning mechanism for speech based on frequency-temporal attention. On the LibriSpeech and TIMIT datasets, the proposed method outperforms the others in terms of area under the curve (AUC), even under extremely low signal-to-noise ratio of -15dB. © 2024 International Speech Communication Association. All rights reserved.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

Improving ECAPA-TDNN Performance with Coordinate Attention

引用

Journal of Shanghai Jiaotong University (science) 2024年 1-7页

作者： Liu, Shuanghong Song, Zhida He, Liang School of Computer Science and Technology Xinjiang University Urumqi830017 China Xinjiang Key Laboratory of Signal Detection and Processing Urumqi830017 China Department of Electronic Engineering and Beijing National Research Center for Information Science and Technology Tsinghua University Beijing100084 China

The current mainstream networks, such as squeeze and excitation residual neural network (SE-ResNet) and emphasized channel attention, propagation and aggregation based time delay neural network (ECAPA-TDNN), enhance the capability of speaker embedding extractors to extract more discriminative speaker embeddings by incorporating squeeze and excitation (SE) attention within the convolutional blocks. However, the SE attention focuses solely on encoding inter-channel information, overlooking the importance of spatial positional information and time-frequency information, which are crucial for the model’s performance. In this paper, we first experimentally compare the effectiveness of several mainstream attention mechanisms in the computer vision domain for the ECAPA-TDNN model. Next, we focus on the substantial improvements that coordinate attention (CA) brings to the ECAPA-TDNN model. The introduction of CA can help the model embed time-frequency information into the channel representation. Even without using AS-Norm, our proposed model achieves relative reductions of about 5.3% equal error rate (EER) and 5.5% minimum detection cost function (minDCF) on both the Voxceleb-O and Voxceleb-H test sets compared to the ECAPA-TDNN baseline model. In addition, the EER is relatively reduced by 9.46% on the CN-Celeb1 test set. This result strongly demonstrates that the CA module can effectively improve the generalization ability of the ECAPA-TDNN model. © Shanghai Jiao Tong University 2024.

关键词： Neural network models

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：