检索结果-内蒙古大学图书馆

DPG: a model to build feature subspace against adversarial patch attack

MACHINE LEARNING 2024年第8期113卷 5601-5622页

作者： Xue, Yunsheng Wen, Mi He, Wei Li, Weiwei Shanghai Univ Elect Power Comp Sci & Technol Hucheng Ring Rd Shanghai 201306 Peoples R China Third Inst Publ Secur Testing Ctr Yueyang Rd Shanghai 200031 Peoples R China

Adversarial patch attacks in the physical world are a major threat to the application of deep learning. However, current research on adversarial patch defense algorithms focuses on image pre-processing defenses, it has been demonstrated that this defense reduces the classification accuracy of clean images and is unable to defend against physically realizable attacks. In this paper, we propose a defense patch GNN (DPG), using a new perspective for defending against adversarial patch attacks. First, we extract the input image features with the feature extraction to obtain a feature set. Then downsampling the feature set by applying the global average pooling layer to reduce the perturbation of the features by the adversarial patch. Finally, this paper proposes a graph-structured feature subspace to robust the feature performance. In addition, we design an optimization algorithm based on stochastic gradient descent (SGD), which can significantly increase the mode's generalization ability. We demonstrate empirically the superior robustness of the DPG model on existing adversarial patch attacks. DPG shows without any accuracy loss in the prediction of clean images.

关键词： Image classification Adversarial defense Feature subspace optimization algorithm

来源：评论

学校读者我要写书评

暂无评论

Sinusoidal Guidance

引用

JOURNAL OF GUIDANCE CONTROL AND DYNAMICS 2024年第3期47卷 417-432页

作者： Kedarisetty, Siddhardha Shima, Tal Technion Israel Inst Technol IL-32000 Hefa Israel

This work presents a novel three-point guidance strategy that simultaneously achieves a desired impact angle, launch angle, impact time, and maximum lateral acceleration against a stationary target. The guidance strategy uses four sinusoidal harmonics of different amplitudes to construct an interceptor trajectory that satisfies all the mission specifications. A new geometric constant G for sinusoids is proposed, which requires only the interceptor's line-of-sight angles from the launcher and the target for computation. Correspondingly, a geometric rule is developed-the interceptor follows the desired sinusoidal trajectory by maintaining G=0 throughout the engagement. A trajectory design algorithm to determine the desired sinusoidal trajectory based on the mission specifications is presented. A proportional-derivative controller is applied to implement the proposed geometric rule, and analytical expressions for the controller gains are defined. The efficacy of the guidance strategy is investigated via various simulation scenarios. It is shown that the developed guidance law is robust to the initial heading errors and interceptor dynamics.

关键词： Guidance, Navigation, and Control Systems Proportional Derivative Controller optimization algorithm Dynamic Modelling Missile Guidance and Control

来源：评论

学校读者我要写书评

暂无评论

Adaptive momentum with discriminative weight for neural network stochastic optimization

引用

INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS 2022年第9期37卷 6531-6554页

作者： Bai, Jiyang Ren, Yuxiang Zhang, Jiawei Florida State Univ Dept Comp Sci Tallahassee FL 32306 USA Florida State Univ Dept Comp Sci IFM Lab Tallahassee FL 32306 USA Univ Calif Davis IFM Lab Dept Comp Sci Davis CA 95616 USA

optimization algorithms with momentum have been widely used for building deep learning models because of the fast convergence rate. Momentum helps accelerate Stochastic gradient descent in relevant directions in parameter updating, minifying the oscillations of the parameters update route. The gradient of each step in optimization algorithms with momentum is calculated by a part of the training samples, so there exists stochasticity, which may bring errors to parameter updates. In this case, momentum placing the influence of the last step to the current step with a fixed weight is obviously inaccurate, which propagates the error and hinders the correction of the current step. Besides, such a hyperparameter can be extremely hard to tune in applications as well. In this paper, we introduce a novel optimization algorithm, namely, Discriminative wEight on Adaptive Momentum (DEAM). Instead of assigning the momentum term weight with a fixed hyperparameter, DEAM proposes to compute the momentum weight automatically based on the discriminative angle. The momentum term weight will be assigned with an appropriate value that configures momentum in the current step. In this way, DEAM involves fewer hyperparameters. DEAM also contains a novel backtrack term, which restricts redundant updates when the correction of the last step is needed. The backtrack term can effectively adapt the learning rate and achieve the anticipatory update as well. Extensive experiments demonstrate that DEAM can achieve a faster convergence rate than the existing optimization algorithms in training the deep learning models of both convex and nonconvex situations.

关键词： deep learning momentum neural network training optimization algorithm stochastic optimization

来源：评论

学校读者我要写书评

暂无评论

Distributed Least Absolute Deviations Estimation

引用

JOURNAL OF GUIDANCE CONTROL AND DYNAMICS 2024年第1期47卷 49-60页

作者： Prabhu, Kaushik Alfriend, Kyle T. Rahmani, Amir Hadaegh, Fred Y. Texas A&M Univ Dept Aerosp Engn College Stn TX 77843 USA CALTECH Dept Aerosp Engn Jet Prop Lab Pasadena CA 91109 USA

Distributed algorithms are essential for reducing communication costs, computational complexity, and memory requirements while performing collaborative estimation using multi-agent systems. Additionally, robustness in estimators is important to prevent performance degradation when the measurement noise is non-Gaussian. Least absolute deviations estimators are known to be robust in the presence of gross errors or outliers in the measurements. To this end, we develop the distributed least absolute deviations (D-LAD) estimator for linear systems whereby the agents iteratively exchange information with their immediate neighbors via single-hop communications to gain a network-wide consensus on the estimates. Additionally, the D-LAD algorithm is implemented in a nonlinear framework to solve the problem of distributed orbit determination of a target body using a formation of spacecraft. Numerical simulations demonstrating the effectiveness of the D-LAD estimator in linear and nonlinear settings are provided.

关键词： Lagrange Multipliers optimization algorithm Distributed Computing Environments Formation Flying Linear System Theory State Estimation

来源：评论

学校读者我要写书评

暂无评论

SOC estimation for lithium-ion battery based on AGA-optimized AUKF

引用

JOURNAL OF ENERGY STORAGE 2024年 75卷

作者： Fan, Xingming Feng, Hao Yun, Xiang Wang, Chao Zhang, Xin Guilin Univ Elect Technol Dept Elect Engn & Automat Guilin 541004 Peoples R China

The window size for covariance matching (CM) of the adaptive unscented Kalman filter (AUKF) affects the state of charge (SOC) estimation performance due to changes with time in the distribution of error innovation sequence (EIS). A new adaptive genetic algorithm (AGA) to address this problem is proposed in this paper. The proposed AUKF (AGA-AUKF1) obtains its best window size determined by the AGA. A novel AUKF (AGA-AUKF2) is proposed to prevent the uncertainty caused by time-varying EIS with the combination of AGA and CM methods. Firstly, the influence of different temperatures on the prediction performance of the algorithm is investigated by FUDS data. The influence of various parameters on the algorithm is further analyzed by FUDS. For different temperatures, initial SOC values, and initial measurement noise covariance, the SOC estimation results show that the accuracy of AGA-AUKFs is better than AUKF. The population size and termination algebra simulation results indicate that the proposed AGA performs well in parameter optimization. Subsequently, the SOC estimation capability of two proposed methods in different working conditions is analyzed by BJDST and US06. The results show that AGA-AUKF2 has better accuracy and robustness than AGA-AUKF1.

关键词： State of charge estimation Adaptive genetic algorithm Adaptive unscented Kalman filter optimization algorithm Covariance matching

来源：评论

学校读者我要写书评

暂无评论

An overview of PSO method for PA optimization

An overview of PSO method for PA optimization

引用

2024 Asia-Pacific Microwave Conference

作者： Jin, Benhao Liu, Yin Cai, Jialin Hangzhou Dianzi Univ Innovat Ctr Elect Design Automat Technol Hangzhou Peoples R China Univ Elect Sci & Technol China Natl Key Lab Wireless Commun Chengdu Peoples R China

ISBN: (纸本)9798350363555;9798350363548

The purpose of this paper is to review the application of particle swarm optimization (PSO) methods to the design of RF power amplifiers (PAs). The basic theory of particle swarm optimization is presented first. Following that, various augmented PSO methods by various techniques will be presented and discussed, including hyperparameter improved techniques, mode-optimized techniques, and integrating PSO with other algorithms. In the final analysis, the PAs obtained by these different optimization algorithms are compared and analyzed.

关键词： EDA tool hyperparameter optimization algorithm particle swarm optimization RF PA

来源：评论

学校读者我要写书评

暂无评论

Development of a novel mathematical model for leakage detection and localization in the water distribution system: based on the modification of the hydraulic model

引用

INTERNATIONAL JOURNAL OF ENVIRONMENTAL SCIENCE AND TECHNOLOGY 2024年第8期21卷 6297-6312页

作者： Rabieian, M. Qaderi, F. Babol Noshirvani Univ Technol Fac Civil Engn Babol Iran

As a result of contamination intrusion's inherent vulnerability, water quality security has been an important issue within water distribution systems (WDSs). In order to detect (un)intentional intrusion events in a timely/effective manner, intensive studies have been undertaken to identify leakage detection and localization methodologies. It is possible to detect and localize leaks in water distribution systems using models-based methodologies. The purpose of this paper is to propose a novel leakage detection and leakage localization model that is based on the network hydraulic model. In order to solve the leakage problem, all hydraulic relationships have been modified and a new model has been developed. This model is developed based on head variation of network sensor nodes. This study utilized mathematical modeling based on the response surface methodology to detect leakages, in addition to detecting leaks;this method can also be used to assess the location of sensors. Consequently, in addition to developing a novel model, a new method is presented for assessing sensor placement in the present study. A leakage diagnosis benchmark dataset is used to demonstrate the proposed methodology and evaluate its effectiveness. Based on the final results, the presented method performed well and was highly accurate.

关键词： optimization algorithm Invasive weed optimization Pipe damage Fault diagnosis Model-based methods Response surface methodology

来源：评论

学校读者我要写书评

暂无评论

Human Conception Optimizer-Based Optimal Type-2 Fuzzy PID Controller Design for Artificial Respiratory System

引用

IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS 2024年第9期54卷 5388-5399页

作者： Acharya, Debasis Das, Dushmanta Kumar Brainware Univ Dept Comp Sci Engn Kolkata 700125 India Natl Inst Technol Nagaland Dept Elect & Elect Engn Dimapur 797103 India

The purpose of this article is to design an optimal fuzzy type-2 proportional integral derivative (FT2PID) controller to enhance the pressure tracking capability of an artificial respiratory system. A patient-hose blower-driven mechanical ventilator (MV) operated in pressure-controlled mode is examined with the proposed controller structure. The error between the desired airway pressure and the ventilator pressure is used as an input to the fuzzy type-2 controller. Another fuzzy input is the change in error. The output variables of the fuzzy inference system (FIS) of the fuzzy controller in the proposed control structure are the parameters of a PID controller. The ranges and points of a triangular-shaped fuzzy type-2 inference system are optimized for the ventilator model with a newly introduced optimizer named the human conception optimizer (HCO) algorithm. With the optimized fuzzy type-2 controller, the parameters of the PID controller are adjusted automatically during any external disturbance or in the presence of any parametric uncertainties in the system. The inherent features of handseling uncertainties of the fuzzy type-2 controller are verified with the PID controller for the ventilator model under different scenarios. With the proposed control scheme, the pressure tracking profile of the ventilator is improved in terms of response time, settling time, and overshoot as compared to the existing results.

关键词： Fuzzy type-2 system mechanical ventilator (MV) optimal membership function (MF) optimization algorithm PID controller

来源：评论

学校读者我要写书评

暂无评论

Desensitized Optimal Control

引用

JOURNAL OF GUIDANCE CONTROL AND DYNAMICS 2024年第8期47卷 1542-1555页

作者： Seywald, Hans Seywald, Kevin L. Analyt Mech Associates Inc Hampton VA 23666 USA KBR Intelligent Syst Res & Dev Support Houston TX 77002 USA

Desensitized optimal control (DOC) enables the formulation of optimal control problems that incorporate the optimal reference solution's sensitivities to state perturbations into the optimization process. Mathematically, this is achieved through the introduction of Lagrange multiplierlike quantities that act as additional state variables and capture the desired sensitivity information. By penalizing user-specified sensitivities, the reference solution becomes easier to control under feedback. Thus, the DOC method can be viewed as a method to improve nonlinear robustness. By converting fixed parameters appearing in the problem formulation to state variables with trivial dynamics, the DOC approach can be used to improve robustness to parameter uncertainties. The paper also analyzes the relationship between sensitivities and the covariance matrix and compares the benefits and limitations of covariance shaping versus DOC. A simple Zermelo-type boat path optimization problem with uncertainties in the water current is analyzed to demonstrate the DOC approach.

关键词： Optimal Control Problem Numerical Integration Lyapunov Equation Feedback Control Additive White Gaussian Noise Engineering Economics Mathematical Analysis optimization algorithm

来源：评论

学校读者我要写书评

暂无评论

Research on Deformation and Failure Mechanism of Surrounding Rock in Inclined Strata with Soft-Hard Interbedded Roadways

引用

ROCK MECHANICS AND ROCK ENGINEERING 2024年第12期57卷 11319-11332页

作者： Ma, Xinmin Zhai, Zhonghua Xiang, Junjie Wei, Mengjie Tian, Zhihui Chen, Liying China Univ Min & Technol Sch Mech & Civil Engn Beijing 100083 Peoples R China

In order to address the challenge of stability control of surrounding rock in roadways under complex geological conditions such as deep high-stress inclined strata and soft-hard interbedding, a comprehensive approach combining similarity model tests and complex variable function analysis with the Adam optimization algorithm was employed. This approach was utilized to obtain the stress and failure characteristics of surrounding rock in roadways under inclined strata and soft-hard interbedding conditions, elucidating the stability failure mechanism of surrounding rock in roadways. The research results indicate that: the internal stress direction in inclined strata shows obvious orientation, tending to be perpendicular to the stratigraphic boundary;the stress distribution in roadway surrounding rock varies with the inclination direction of the strata. Specifically, when the strata inclination angle is clockwise, stress concentration tends to occur on the left sidewall and right floor of the roadway;conversely, when the strata inclination angle is counterclockwise, stress concentration tends to occur on the right sidewall and left floor of the roadway. The use of the Adam optimization algorithm for solving mapping functions demonstrates superiority, with a solution time of only 1.25 s and high accuracy. When the number of coefficient terms in the mapping function is 9, the average error is only 0.17%, which is 0.34 times lower compared to other optimization methods, with the required number of iterations reduced by 0.64, significantly reducing subsequent computational pressure. For cross-layer roadways with soft-hard interbedding surrounding rock, the deformation and failure of roadways are closely related to the interface position between the rock layers, with deformation more likely to occur in the soft rock area and with larger deformations. Under conditions of high stress, when facing both the soft-hard interbedding and inclined strata issues simultaneously, priorit

关键词： Inclined strata Soft-hard interbedding Model test Complex variable function optimization algorithm

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：