检索结果-内蒙古大学图书馆

The Journal of Machine Learning Research 2023年第1期24卷 14474-14523页

作者： Haishan Ye Luo Luo Ziang Zhou Tong Zhang Center for Intelligent Decision-Making and Machine Learning School of Management Xi'an Jiaotong University Xi'an China School of Data Science Fudan University Shanghai China Department of Computing The Hong Kong Polytechnic University Hong Kong China Computer Science & Mathematics The Hong Kong University of Science and Technology Hong Kong China

This paper considers the decentralized convex optimization problem, which has a wide range of applications in large-scale machine learning, sensor networks, and control theory. We propose novel algorithms that achieve optimal computation complexity and near optimal communication complexity. Our theoretical results give affirmative answers to the open problem on whether there exists an algorithm that can achieve a communication complexity (nearly) matching the lower bound depending on the global condition number instead of the local one. Furthermore, the linear convergence of our algorithms only depends on the strong convexity of global objective and it does not require the local functions to be convex. The design of our methods relies on a novel integration of well-known techniques including Nesterov's acceleration, multi-consensus and gradient-tracking. Empirical studies show the outperformance of our methods for machine learning applications.

关键词： consensus optimization decentralized algorithm accelerated gradient descent gradient tracking composite optimization

来源：评论

学校读者我要写书评

暂无评论

SoVAR: Building Generalizable Scenarios from Accident Reports for Autonomous Driving Testing

arXiv

引用

arXiv 2024年

作者： Guo, An Zhou, Yuan Tian, Haoxiang Fang, Chunrong Sun, Yunjian Sun, Weisong Gao, Xinyu Luu, Anh Tuan Liu, Yang Chen, Zhenyu State Key Laboratory for Novel Software Technology Nanjing University Nanjing210023 China School of Computer Science and Technology Zhejiang Sci-Tech University Hangzhou310018 China College of Computing and Data Science Nanyang Technological University Singapore

Autonomous driving systems (ADSs) have undergone remarkable development and are increasingly employed in safety-critical applications. However, recently reported data on fatal accidents involving ADSs suggests that the desired level of safety has not yet been fully achieved. Consequently, there is a growing need for more comprehensive and targeted testing approaches to ensure safe driving. Scenarios from real-world accident reports provide valuable resources for ADS testing, including critical scenarios and high-quality seeds. However, existing scenario reconstruction methods from accident reports often exhibit limited accuracy in information extraction. Moreover, due to the diversity and complexity of road environments, matching current accident information with the simulation map data for reconstruction poses significant challenges. In this paper, we design and implement SoVAR, a tool for automatically generating road-generalizable scenarios from accident reports. SoVAR utilizes well-designed prompts with linguistic patterns to guide the large language model (LLM) in extracting accident information from textual data. Subsequently, it formulates and solves accident-related constraints in conjunction with the extracted accident information to generate accident trajectories. Finally, SoVAR reconstructs accident scenarios on various map structures and converts them into test scenarios to evaluate its capability to detect defects in industrial ADSs. We experiment with SoVAR, using the accident reports from the National Highway Traffic Safety Administration's (NHTSA) database to generate test scenarios for the industrial-grade ADS Apollo. The experimental findings demonstrate that SoVAR can effectively generate generalized accident scenarios across different road structures. Furthermore, the results confirm that SoVAR identified 5 distinct safety violation types that contributed to the crash of Baidu Apollo. Copyright © 2024, The Authors. All rights reserved.

关键词： Automatic test pattern generation

来源：评论

学校读者我要写书评

暂无评论

Towards Federated Learning on Fresh datasets

Towards Federated Learning on Fresh Datasets

引用

IEEE Internatonal Conference on Mobile Adhoc and Sensor Systems (MASS)

作者： Chen Wu Mingjun Xiao Jie Wu Yin Xu Jinrui Zhou He Sun School of Data Science University of Science and Technology of China China School of Computer Science and Technology & Suzhou Institute for Advanced Research University of Science and Technology of China China Department of Computer and Information Sciences Temple University USA

Federated Learning (FL) is an emerging privacy-preserving distributed computing paradigm that enables numerous clients to collaboratively train machine learning models without the need for transmitting the private datasets of clients to the FL server. Unlike most existing research where the local datasets of clients are considered to be unchanging over time during the whole FL process, we consider such scenarios in this paper where the local datasets of clients need to be updated periodically, and the server can stimulate clients to use as fresh as possible datasets to train their local models. Our objective is to determine a client selection strategy to minimize the loss of global model for FL with a limited budget. To this end, we leverage the concept of Age of Information (AoI) to quantify the freshness of local datasets and theoretically analyze the convergence bound of our AoI-aware FL system. Based on the convergence bound, we formalize our problem as a restless multi-armed bandit problem. Then, we devise a Whittle’s-Index-based Client Selection algorithm, called WICS, to tackle the client selection problem. Extensive simulations show that the proposed algorithm can reduce the training loss and improve learning accuracy compared to state-of-the-art algorithms.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Cyber intrusion detection using dual interactive Wasserstein generative adversarial network with war strategy optimization in wireless sensor networks

引用

Multimedia Tools and Applications 2025年第18期84卷 19223-19253页

作者： Anusha, N. Tapas Bapu, B.R. S, Selvakumaran Vijayaraj, A. Ramesh Kumar, C. P, Raji Department of IoT School of Computer science and Engineering VIT Tamil Nadu Vellore India Department of Electronics and Communication Engineering S.A.Engineering College Tamil Nadu Chennai India Department of Artificial Intelligence and Data Science Rajalakshmi Institute of Technology Tamil Nadu Chennai India Department of Information Technology R.M.K Engineering College R.S.M Nagar Tamil Nadu Kavaraipettai India Department of Electronics and Communication Engineering Panimalar Engineering College Tamil Nadu Chennai India Department of Electronics and Communication Engineering Saveetha Engineering College Tamil Nadu Chennai India

Wireless sensor network (WSN) is one of the essential components of a multi-hop cyber-physical system comprising many fixed or moving sensors. There are many common attacks in WSN, which can quickly harm a WSN system. In this manuscript, Dual Interactive Wasserstein Generative Adversarial Network (DIWGAN) with war strategy optimization (WSO) algorithm is introduced for detecting intrusion in WSN. Initially, the input data is attained from the NSL-KDD and CICIDS2017 datasets. The data is then passed onto Morphological filtering and Extended Empirical wavelet transformation (MFEEWT) based pre-processing. After that, the filtered data is sent to the DIWGAN, which detects the intrusion and classifies the fraud data in the network. Additionally, the WSO algorithm is introduced to enhance the classification parameters of DIWGAN. By then the performance of the proposed DIWGAN-WSO-IDS methodology is evaluated using Python Platform. Also, the performance of the proposed methodology is calculated using various metrics like accuracy, precision, recall, f-measure, mean square error, and attack detection rate. Thus, the proposed technique has attained 29.87%, 31.7%, and 26.8% higher accuracy, 23.54%, 20.09%, and 26.98% higher recall, 12.76%, 10.96%, and 6.7% higher precision. And 6.85%, 4.56%, and 8.56% lower MSE than the existing techniques like BFSGSRF, MHICA-SSA, and CNN-LSTM methods. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Wireless sensor networks

来源：评论

学校读者我要写书评

暂无评论

An Efficient Implementation for Bayesian Manifold Regularization Method

An Efficient Implementation for Bayesian Manifold Regulariza...

引用

IEEE Conference on Decision and Control

作者： Junpeng Zhang Yue Ju Bo Wahlberg Biqiang Mu Tianshi Chen School of Data Science and Shenzhen Research Institute of Big Data The Chinese University of Hong Kong Shenzhen China Division of Decision and Control Systems School of Electrical Engineering and Computer Science KTH Royal Instituite of Technology Stockholm Sweden Key Laboratory of Systems and Control Institute of Systems Science Academy of Mathematics and System Science Chinese Academy of Sciences Beijing China

When applying the Bayesian manifold regularization method to function estimation problem with manifold constraints, the direct implementation has computational complexity $\mathcal{O}(N^{3})$ , where $N$ is the number of input-output data measurements. This becomes particularly costly when $N$ is large. In this paper, we propose a more efficient implementation based on the Kalman filter and smoother using a state-space model realization of the underlying Gaussian process. Moreover, we explore the sequentially semi-separable structure of the Laplacian matrix and the posterior covariance matrix. Our proposed implementation has computational complexity $\mathcal{O}(N)$ and thus can be applied to large data problems. We exemplify the effectiveness of our proposed implementation through numerical simulations.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Label Distribution Feature Selection Based on Neighborhood Rough Set

SSRN

引用

SSRN 2024年

作者： Wu, Yilin Guo, Wenzhong Lin, Yaojin College of Computer and Data Science Fuzhou University Fuzhou350116 China Key Laboratory of Network Computing and Intelligent Information Processing Fuzhou University Fuzhou350116 China School of Computer Science Minnan Normal University Zhangzhou363000 China

In label distribution learning, an instance is involved with many labels in different importance degrees, and the feature space of instances is accompanied with thousands of redundant and/or irrelevant features. Therefore, the main characteristic of feature selection in label distribution learning is to evaluate the discriminable ability of each feature. Motivated by neighborhood rough set, which can be used to measure the dependency degree of feature via constructing neighborhood relations on feature space and label space, respectively, this paper proposes a novel lable distribution feature selection method. In this paper, the neighborhood class of instance in label distribution space is defined, which is beneficial to recognize the logical class of target instance. Then, a new neighborhood rough set model for label distribution learning is proposed. Specially, the dependency degree of feature combining label weight is defined. Finally, a label distribution feature selection based on neighborhood rough set is presented. Extensive experiments on 12 data sets show the effectiveness of the proposed algorithm. © 2024, The Authors. All rights reserved.

关键词： Feature Selection

来源：评论

学校读者我要写书评

暂无评论

High-Fidelity Editable Portrait Synthesis with 3D GAN Inversion

High-Fidelity Editable Portrait Synthesis with 3D GAN Invers...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Jindong Xie Jiachen Liu Yupei Lin Jinbao Wang Xianxu Hou Linlin Shen Computer Vision Institute School of Computer Science and Software Engineering Shenzhen University National Engineering Laboratory for Big Data System Computing Technology Shenzhen University School of Information Engineering Guangdong University of Technology Guangzhou China Guangdong Provincial Key Laboratory of Intelligent Information Processing School of AI and Advanced Computing Xi’an Jiaotong-Liverpool University

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

The 3D generative adversarial network (GAN) inversion converts an image into 3D representation to attain high-fidelity reconstruction and facilitate realistic image manipulation within the 3D latent space. However, previous approaches face challenges regarding the trade-off between the reconstruction ability and editability. That is, reversing a real-world image to a low-dimensional latent code would inevitably lead to information loss, and achieving a near-perfect reconstruction using high-rate triplane representation often limits the ability to manipulate the image freely in the latent space. To address these issues, we propose a novel latent conditioning encoder-based framework with the alignment between the low-dimensional latent and high-dimensional triplane. A non-semantic guided editing strategy bridges the intrinsic relation between the latent condition and triplane generation, making it possible to edit the high-dimensional representation by latent manipulation. As a result, our method can achieve high-fidelity reconstruction and editing simultaneously by directly controlling the latent code. Experimental results demonstrate that our approach excels in reconstruction and editing quality compared to previous 3D inversion methods. Furthermore, our method can also edit even real faces with large poses and out-of-domain cases.

关键词： Bridges Three-dimensional displays Codes Signal processing Generative adversarial networks Vectors Acoustics Speech processing Image reconstruction Faces

来源：评论

学校读者我要写书评

暂无评论

Exploiting Meta-Learned Confidences for Imbalanced Multilabel Learning

引用

IEEE Transactions on Neural Networks and Learning Systems 2024年第6期36卷 10242-10256页

作者： Zhihan Ning Zhixing Jiang David Zhang School of Science and Engineering The Chinese University of Hong Kong Shenzhen Guangdong China Institute of Oceanographic Instrumentation Qilu University of Technology (Shandong Academy of Sciences) Qingdao China School of Data Science The Chinese University of Hong Kong Shenzhen Guangdong China Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen Guangdong China

Multilabel learning deals with datasets where each sample is associated with multiple labels. It is commonly assumed that label correlations should be well exploited to build an effective multilabel classifier. Moreover, the class imbalance problem occurs in many multilabel datasets and should be tackled to reduce the classification bias. While many multilabel learning methods have been proposed, research on imbalanced multilabel learning (IMLL) is relatively deficient. To address these issues, we exploit the value of meta-learned confidences, i.e., the prediction confidences iteratively updated over the out-of-bag samples, for IMLL. First, such meta-confidences can be fused to the original feature space to learn high-order label correlations. Second, meta-confidences can be used to calibrate the prediction results to alleviate class imbalance. Motivated by these, we propose an ensemble learning method named meta-confidence ensemble (MCE) for IMLL. Specifically, MCE iteratively makes bootstrap replicates of the multilabel training set, leverages the out-of-bag samples to generate meta-confidences, and fuses them to the original feature space to learn label correlations. A sparse projection method is presented to avoid overfitting and improve the ensemble diversity. Finally, the prediction result of an unseen sample is determined by the calibrated plurality vote of MCE’s base classifiers. Extensive experiments demonstrated the effectiveness and superiority of MCE for IMLL. Codes have been made publicly available at https://***/ CUHKSZ-NING/MCEClassifier.

关键词： Correlation Training Ensemble learning Prediction algorithms Overfitting Learning systems data models Artificial neural networks Aerospace electronics Stacking

来源：评论

学校读者我要写书评

暂无评论

GENERATIVE ADVERSARIAL NETWORKS FOR SPATIO-SPECTRAL COMPRESSION OF HYPERSPECTRAL IMAGES

arXiv

引用

arXiv 2023年

作者： Fuchs, Martin Hermann Paul Byju, Akshara Preethy Walda, Alisa Rasti, Behnood Demir, Begüm Faculty of Electrical Engineering and Computer Science Technische Universität Berlin Germany Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Amritapuri India BIFOLD - Berlin Institute for the Foundations of Learning and Data Germany

Deep learning-based hyperspectral image (HSI) compression has recently attracted great attention in remote sensing due to the growth of hyperspectral data archives. Most of the existing models achieve either spectral or spatial compression and do not jointly consider the spatio-spectral redundancies present in HSIs. To address this problem, in this paper, we propose High Fidelity Compression (HiFiC)-based models for spatio-spectral compression of HSIs. In detail, we introduce two new models: i) HiFiC using Squeeze and Excitation (SE) blocks (denoted as HiFiCSE);and ii) HiFiC with 3D convolutions (denoted as HiFiC3D) in the framework of compression of HSIs. We analyze the effectiveness of HiFiCSE and HiFiC3D in compressing the spatio-spectral redundancies with channel attention and inter-dependency analysis. Experimental results show the efficacy of the proposed models in performing spatio-spectral compression, while reconstructing images at reduced bitrates with higher reconstruction quality. The code of the proposed models is publicly available at https://***/rsim/HSI-SSC. © 2023, CC BY.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

MA-Unet3+ Segmentation Network of Remote Sensing Image Based on ECA Block

MA-Unet3+ Segmentation Network of Remote Sensing Image Based...

引用

International Conference on Computer Network, Electronic and Automation (ICCNEA)

作者： Lang Gao Jianguo Wang Xin Ye School of Weapon Science and Technology Xi'an Technological University Xi’an China Research Institute of Artificial Intelligence and Data Science Xi'an Technological University Xi’an China

Semantic segmentation is a basal task and is a typical computer vision problem. Although semantic segmentation is developing rapidly, the speed and accuracy of model segmentation still need to be further improved. For solve the issue of scale differences between target objects and loss of spatial information in the segmentation task of remote sensing images, by improving the original U-Net3+ network and introducing the attention mechanism, a new network MA-Unet3+ is constructed. In the coding phase, images of unlike scales are fused, and the full-scale connections are pruned, some skip connections are removed, and attention mechanisms are introduced between each layer. The improved model is contrast with some common network models, and the experiment achieves 78.7% average intersection (mIoU) on the Vaihingen dataset, which is 0.8% better than this optimized network U-Net3+, the average category pixel accuracy (MPA) is 92.4%, which is 1.2% better, and the similarity coefficient (Dice) result is 87.3%, which is 0.8% better. 0.8%, it is observed that MA-Unet3+ is precede other algorithms.

关键词：

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：