检索结果-内蒙古大学图书馆

FedQP: Large-Scale Private and Flexible Federated Query Processing

IEEE Access 2025年 13卷 92019-92035页

作者： Almohri, Hussain M. J. Watson, Layne T. Kuwait University Department of Computer Science Kuwait City Kuwait Virginia Tech Departments of Computer Science Mathematics Aerospace and Ocean Engineering BlacksburgVA United States

State-of-the-art federated learning methods orchestrate iterations of the stochastic gradient descent algorithm among a network of clients to refine a unified set of model parameters, all while safeguarding individual client datasets. Although promising, the success of existing approaches hinges on using a uniform data model with the same optimization and learning algorithms by the participants. In addition, existing methods can be subject to privacy attacks, such as model-inversion attacks. A pivotal challenge lies in crafting an algorithm that can efficiently aggregate client inputs, ensuring minimal client dataset exposure and maximum flexibility in model adaptation. We propose a new scheme for aggregating local, diverse, and independent data models, addressing data inconsistency and model-inversion attacks. The proposed scheme, federated query processing FedQP, enables clients to build local models independently, without coordinating with the server or other clients. The server communicates with selected clients only to predict future values. Our scheme only requires one communication round between each client and the server for model initialization without the need for a training phase. Clients do not share data or model parameters. We evaluated FedQP by assessing the impact of our new aggregation mechanism on model accuracy and the communication efficacy of our architecture. We show that the caching algorithm in FedQP reduces redundant client communications during query processing. In addition, we present a privacy analysis that shows that our method outperforms prominent gradient-based approaches to federated learning. Our experiments show that FedQP consistently achieves higher classification accuracy in non-IID settings and demonstrates stronger resilience to reconstruction attacks compared to gradient-based methods such as federated averaging (FedAvg). © 2013 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Zero-dynamics attack detection based on data association in feedback pathway

引用

Cognitive Robotics 2025年 5卷 126-139页

作者： Zhang, Zeyu Li, Hongran Todo, Yuki Division of Electrical Engineering and Computer Science Kanazawa University Ishikawa Kanazawa9201192 Japan School of Computer Engineering Jiangsu Ocean University Jiangsu Lianyungang222005 China Faculty of Electrical Information and Communication Engineering Kanazawa University Ishikawa Kanazawa9201192 Japan

This paper considers the security of non-minimum phase systems, a typical kind of cyber-physical systems. Non-minimum phase systems are characterized by unstable zeros in their transfer functions, making them particularly susceptible to disturbances and attacks. The non-minimum phase systems are more vulnerable to zero-dynamics attack (ZDA) than minimum phase systems. ZDA is a stealthy attack strategy that exploits the internal dynamics of a system, remaining undetectable while causing gradual system destabilization. Recent cyber incidents have demonstrated the increasing risk of such hidden attacks in critical infrastructures, such as power grids and transportation systems. This paper first demonstrates that the existing ZDA has the limitation of falling into local convergence, and then proposes an enhanced zero-dynamics attack (EZDA), which overcomes local convergence by diverging system data. Furthermore, this paper presents an autoregressive model which can build the data association between the original data and the forged data. By observing the fluctuations in state values, the presented model can detect not only ZDA, but also EZDA. Finally, numerical simulations and an application example are provided to verify the theoretical results. © 2025

关键词： Cyber attacks

来源：评论

学校读者我要写书评

暂无评论

Reconfigurable MIMO-based self-powered battery-less light communication system

引用

Light(science & Applications) 2024年第10期13卷 2300-2311页

作者： Jose Ilton De Oliveira Filho Abderrahmen Trichili Omar Alkhazragi Mohamed-Slim Alouini Boon S.Ooi Khaled Nabil Salama Computer Electrical and Mathematical Sciences&EngineeringKing Abdullah University of Science and TechnologyThuwalSaudi Arabia NEOM Institute for Ocean Science and Solutions NeomThuwalSaudi Arabia

Simultaneous lightwave information and power transfer (SLIPT), co-existing with optical wireless communication,holds an enormous potential to provide continuous charging to remote Internet of Things (IoT) devices while ensuringconnectivity. Combining SLIPT with an omnidirectional receiver, we can leverage a higher power budget whilemaintaining a stable connection, a major challenge for optical wireless communication systems. Here, we design amultiplexed SLIPT-based system comprising an array of photodetectors (PDs) arranged in a 3 × 3 configuration. Thesystem enables decoding information from multiple light beams while simultaneously harvesting energy. The PDs canswiftly switch between photoconductive and photovoltaic modes to maximize information transfer rates and provideon-demand energy harvesting. Additionally, we investigated the ability to decode information and harvest energywith a particular quadrant set of PDs from the array, allowing beam tracking and spatial diversity. The design wasexplored in a smaller version for higher data rates and a bigger one for higher power harvesting. We report a selfpoweringdevice that can achieve a gross data rate of 25.7 Mbps from a single-input single-output (SISO) and an 85.2Mbps net data rate in a multiple-input multiple-output (MIMO) configuration. Under a standard AMT1.5 illumination,the device can harvest up to 87.33 mW, around twice the power needed to maintain the entire system. Our workpaves the way for deploying autonomous IoT devices in harsh environments and their potential use in spaceapplications.

关键词： communication MIMO system

来源：评论

学校读者我要写书评

暂无评论

Robust Conformal Prediction Using Privileged Information 38

Robust Conformal Prediction Using Privileged Information

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Feldman, Shai Romano, Yaniv Department of Computer Science Technion Israel Departments of Electrical and Computer Engineering and of Computer Science Technion Israel

We develop a method to generate prediction sets with a guaranteed coverage rate that is robust to corruptions in the training data, such as missing or noisy variables. Our approach builds on conformal prediction, a powerful framework to construct prediction sets that are valid under the i.i.d assumption. Importantly, naively applying conformal prediction does not provide reliable predictions in this setting, due to the distribution shift induced by the corruptions. To account for the distribution shift, we assume access to privileged information (PI). The PI is formulated as additional features that explain the distribution shift, however, they are only available during training and absent at test time. We approach this problem by introducing a novel generalization of weighted conformal prediction and support our method with theoretical coverage guarantees. Empirical experiments on both real and synthetic datasets indicate that our approach achieves a valid coverage rate and constructs more informative predictions compared to existing methods, which are not supported by theoretical guarantees. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Centripetal Intensive Deep Hashing for Remote Sensing Image Retrieval

引用

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 2025年 18卷 12439-12453页

作者： Wang, Weigang Guo, Zhongwen Cui, Ziyuan Zhao, Hailei Xian, Lintao The Faculty of Electrical Engineering and Computer Science Ocean University of China Qingdao266499 China The Department of Intelligent Medical Engineering Shandong Second Medical University Weifang261053 China

With the breakthrough of convolutional neural networks, deep hashing methods have demonstrated remarkable performance in large-scale image retrieval tasks. However, existing deep supervised hashing methods, which rely on pairwise or triplet labels, typically learn the hash function via random or hardest sample mining within training batches. This strategy primarily captures local sample similarities, causing a distribution shift and limiting retrieval performance. Furthermore, most methods emphasize global features while overlooking structural information, which is essential for understanding spatial relationships in images. To solve these limitations, we propose a Centripetal Intensive Deep Hashing (CIDH) method for remote sensing image retrieval. Initially, we design a Hybrid-Attention Guided Multiscale Refinement Network (HGMRN) that integrates channel and spatial attention to capture multiscale visual features and highlight salient regions at different scales. Subsequently, we introduce a central similarity loss via class-centered labels to optimize the spatial distribution of global samples, which can encourage hash codes with similar semantics to cluster around centroids and reduce distribution shift. Meanwhile, we incorporate a central intensive loss into the Hamming space to shorten intra-class Hamming distances, generating more compact and discriminative hash codes. Extensive experiments demonstrate the superiority of our CIDH method compared with current state-of-the-art deep hashing methods. © 2008-2012 IEEE.

关键词： Hamming distance

来源：评论

学校读者我要写书评

暂无评论

A Partial-Givens-Rotation-Based Symbol Detector for GSM MIMO Systems: Algorithm and VLSI Implementation

引用

IEEE Systems Journal 2023年第4期17卷 6322-6333页

作者： Yen, Mao-Hsu Lu, Hoang-Yang Lin, Shao-Yueh Lu, Ken-Hua Chan, Chia-Chen National Taiwan Ocean University Department of Computer Science Engineering Keelung202 Taiwan National Taiwan Ocean University Department of Electrical Engineering Keelung202 Taiwan

Recently, generalized spatial modulation (GSM) multiple-input multiple-output (MIMO) systems have attracted intensive research interest due to their advantages in balancing spectral efficiency and interchannel interference. Aiming at satisfactory detection performance based on a feasible hardware architecture, in this paper, a partial Givens rotation (PGR)-based symbol detector with a very-large-scale integration (VLSI) hardware architecture is proposed for GSM MIMO systems. The proposed detector contains three main types of modules: PGR blocks, symbol estimation (SE), and minimization. In particular, compared to conventional Givens rotations (GRs), the proposed PGR mechanism can further reduce computational complexity by at least 36 %. In addition, to ensure numerical stability, only adders, shifters, and comparators are used to implement the SE architecture, avoiding the use of dividers. Furthermore, instead of the 2-norm distance measure, the 1-norm distance measure is used in the proposed detector to reduce the number of multipliers, thereby accelerating the detection speed. Finally, computer simulations showthat the proposed algorithm performs achieves near-optimal performance while incurring a lower computational complexity. Additionally, hardware implementation results achieved in TSMC 90-nm CMOS technology at an operating frequency of 704.2 MHz, with a configuration of four transmit antennas, two active transmit antennas, four receive antennas, and 16-ary quadrature amplitude modulation (16-QAM), show that the proposed hardware architecture needs 395.2 k gates and provides a detection throughput of 2347 Mbps and a hardware efficiency of 5.94 Mbps/kGEs for fast fading channels. In comparison to existing works, the proposed detector provides attractive detection performance as well as a feasible hardware architecture. © 2007-2012 IEEE.

关键词： Spectrum efficiency

来源：评论

学校读者我要写书评

暂无评论

The development of assisted- visually impaired people robot in the indoor environment based on deep learning

引用

Multimedia Tools and Applications 2024年第3期83卷 6555-6578页

作者： Hsieh, Yi-Zeng Ku, Xiang-Long Lin, Shih-Syun Department of Electrical Engineering National Taiwan University of Science and Technology Taipei City106 Taiwan Department of Computer Science and Engineering National Taiwan Ocean University Keelung City Taiwan

The indoor positioning for visually impaired people has influence on their daily life in unknown indoor environment. This study designs the robot that can assist the blind walking safety and navigate in indoor environment by a single camera. The sense classification is proposed to position the blind in indoor environment by proposed convolutional neural network framework and integrate the semantic segmentation to find the road surface through a depth camera to guide the blind walking. The proposed vision-based sense classification method is compared with the traditional WiFi triangular-positioning method, and the average error of x-y coordinate position result as (9.25,3.65) is better. From the experiment, the designed robot can help the visually impaired people to indoor navigation in unknown indoor environment. © 2023, The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature.

关键词： Robots

来源：评论

学校读者我要写书评

暂无评论

Deep Convolutional Generative Adversarial Network for Inverse Kinematics of Self-Assembly Robotic Arm Based on the Depth Sensor

引用

IEEE Sensors Journal 2023年第1期23卷 758-765页

作者： Hsieh, Yi-Zeng Xu, Fu-Xiong Lin, Shih-Syun National Taiwan University of Science and Technology Department of Electrical Engineering Taipei City106335 Taiwan National Taiwan Ocean University Department of Computer Science and Engineering Keelung City202301 Taiwan

In this study, we propose a new deep convolutional generative adversarial kinematics network (DCGAKN) to establish inverse kinematics of self-assembly robotic arm. We design that the robot system uses a depth sensor detecting an object by you only look once v4 (YOLOv4) algorithm, and then, our proposed DCGAKN is with generator and discriminator of adversarial evolution training inverse kinematics model for controlling self-assembly robotic arm to solve the limited solution space to be more adaptive in the dynamic environment. The following are advancements of our proposed method: 1) generator neural network is trained by few-shot training data to control the self-assembly robotic arm to achieve high-accuracy position;2) generator is evaluated with discriminator not only depending on training data but also on adaptive evolutionary;3) the self-assembly robotic arm is like humanoid arm not traditional robotic arm structure and it is easy for self-assembly model to build inverse kinematics without computing inverse kinematics matrix;4) the object is detected by the depth information based on YOLOv4;and 5) through generator evolutionary, the activity range of robotic arm is not limited with training data range. The proposed DCGAKN is compared with convolutional neural network (CNN) and deep neural network (DNN) that the accuracy rate and distance error achieve 87% and 1.26 cm, respectively. The source code of this work is at: https://***/YiZengHsieh/DCGAKN. © 2001-2012 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Improved Trade-Offs Between Amortization and Download Bandwidth for Linear HSS 5

Improved Trade-Offs Between Amortization and Download Bandwi...

引用

5th Conference on Information-Theoretic Cryptography, ITC 2024

作者： Blackwell, Keller Wootters, Mary Department of Computer Science Stanford University CA United States Departments of Computer Science and Electrical Engineering Stanford University CA United States

ISBN: (纸本)9783959773331

A Homomorphic Secret Sharing (HSS) scheme is a secret-sharing scheme that shares a secret x among s servers, and additionally allows an output client to reconstruct some function f(x) using information that can be locally computed by each server. A key parameter in HSS schemes is download rate, which quantifies how much information the output client needs to download from the servers. Often, download rate is improved by amortizing over instances of the problem, making also a key parameter of interest. Recent work [23] established a limit on the download rate of linear HSS schemes for computing low-degree polynomials and constructed schemes that achieve this optimal download rate;their schemes required amortization over = Ω(s log(s)) instances of the problem. Subsequent work [6] completely characterized linear HSS schemes that achieve optimal download rate in terms of a coding-theoretic notion termed optimal labelweight codes. A consequence of this characterization was that = Ω(s log(s)) is in fact necessary to achieve optimal download rate. In this paper, we characterize all linear HSS schemes, showing that schemes of any download rate are equivalent to a generalization of optimal labelweight codes. This equivalence is constructive and provides a way to obtain an explicit linear HSS scheme from any linear code. Using this characterization, we present explicit linear HSS schemes with slightly sub-optimal rate but with much improved amortization = O(s). Our constructions are based on algebraic geometry codes (specifically Hermitian codes and Goppa codes). © Keller Blackwell and Mary Wootters.

关键词： Forward error correction

来源：评论

学校读者我要写书评

暂无评论

A Single-Weight Neural Network for MIMO Signal Detection: Algorithm and Implementation 10

A Single-Weight Neural Network for MIMO Signal Detection: Al...

引用

10th International Conference on engineering and Emerging Technologies, ICEET 2024

作者： Ho, Yi-Cheng Yen, Mao-Hsu Lu, Hoang-Yang National Taiwan Ocean University Dept. of Computer Science and Engineering Keelung Taiwan National Taiwan Ocean University Dept. of Electrical and Engineering Keelung Taiwan

ISBN: (纸本)9798331532895

Existing deep learning-based MIMO detectors often face substantial memory demands due to the need to store neural network weights, posing a significant challenge to practical implementation. To address this issue, we propose a single-weight neural network, namely SW-Net. By reducing each layer's weight vector to a scalar, the required hardware storage is significantly reduced. Simulation results demonstrate that hardware requirements can be reduced while maintaining bit error rate (BER) performance. Additionally, we present the corresponding VLSI hardware architecture, implemented using TSMC's 90-nm CMOS technology, operating at 707 MHz. This design requires 1109 kGEs and achieves a throughput of 707 Mbps. © 2024 IEEE.

关键词： Deep neural networks

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：