检索结果-内蒙古大学图书馆

45th IEEE Symposium on Security and Privacy (SP)

作者： Sandeepa, Chamara Siniarski, Bartlomiej Wang, Shen Liyanage, Madhusanka Univ Coll Dublin Sch Comp Sci Dublin Ireland

ISBN: (纸本)9798350331318;9798350331301

With the rapid progression of communication and localisation of big data over billions of devices, distributed Machine Learning (ML) techniques are emerging to cater for the development of Artificial Intelligence (AI)-based services in a distributed manner. Federated Learning (FL) is such an innovative approach to achieve a privacy-preserved AI that facilitates ML model sharing and aggregation while keeping the participants' data at the original source. However, recent research has investigated threats from poisoning attacks in FL. Several robust algorithms based on techniques such as similarity metrics or anomaly filtering are proposed as solutions. Yet, these approaches do not focus on investigating the intentions of the attackers or providing justifications and evidence for suspecting the behaviour of clients who are considered poisoners. Therefore, we propose SHERPA, a robust algorithm that uses Shapley Additive Explanations (SHAP) to identify potential poisoners in an FL system. Based on this, we develop a novel algorithm to differentiate poisoners via feature attribution clustering. We launch data poisoning attacks for different scenarios on multiple datasets and showcase our solution to mitigate the attacks. Furthermore, we show that privacy-targeted poisoning attacks can be mitigated with our approach. Accompanying the Explainable AI (XAI) technique for defence, our study reveals the potential for post-hoc feature attributions in countering data poisoning attacks with better explainability and improved justification in eliminating potentially malicious clients in the aggregation process.

关键词： Federated Learning XAI SHAP Data Poisoning Beyond 5G/6G Privacy Defence robust algorithm

来源：评论

学校读者我要写书评

暂无评论

robust Rendezvous Planning Under Maneuver Execution Errors

引用

JOURNAL OF GUIDANCE CONTROL AND DYNAMICS 2015年第1期38卷 76-93页

作者： Louembet, Christophe Arzelier, Denis Deaconu, Georgia CNRS Lab Anal & Architectures Syst F-31400 Toulouse France

The problem of designing a rendezvous guidance maneuver plan robust to thrusting errors is addressed in this paper. The aim of this paper is to develop tractable and robust guidance algorithms. Solving the rendezvous guidance problem via a direct approach leads to uncertain optimization problems while accounting for the guidance, navigation, and control systems uncertainties and errors. A worst-case approach is considered in order to obtain tractable robust counterparts. The robustness certificates derived from these guidance algorithms provide the means to analyze the effects of the considered errors on the rendezvous mission. Several types of missions tested in a linear environment are used to illustrate the methodology.

关键词： Rendezvous Guidance robust Guidance algorithm GNC Systems Linear Quadratic Regulator Pontryagin's Maximum Principle robust algorithm Numerical algorithms PROBA Nonlinear Dynamics Spacecraft Rendezvous

来源：评论

学校读者我要写书评

暂无评论

robust metric learning based on the rescaled hinge loss

引用

INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS 2020年第11期11卷 2515-2528页

作者： Al-Obaidi, Sumia Abdulhussien Razooqi Zabihzadeh, Davood Hajiabadi, Hamideh Minist Higher Educ & Sci Res Supervis & Sci Evaluat Apparat Baghdad Iraq Sabzevar Univ New Technol Engn Fac Comp Dept Sabzevar Iran Birjand Univ Technol Dept Comp Engn Birjand Iran

Distance/Similarity learning is a fundamental problem in machine learning. For example, kNN classifier or clustering methods are based on a distance/similarity measure. Metric learning algorithms enhance the efficiency of these methods by learning an optimal distance function from data. Most metric learning methods need training information in the form of pair or triplet sets. Nowadays, this training information often is obtained from the Internet via crowdsourcing methods. Therefore, this information may contain label noise or outliers leading to the poor performance of the learned metric. It is even possible that the learned metric functions perform worse than the general metrics such as Euclidean distance. To address this challenge, this paper presents a new robust metric learning method based on the Rescaled Hinge loss. This loss function is a general case of the popular Hinge loss and initially introduced in Xu et al. (Pattern Recogn 63:139-148, 2017) to develop a new robust SVM algorithm. In this paper, we formulate the metric learning problem using the Rescaled Hinge loss function and then develop an efficient algorithm based on HQ (Half-Quadratic) to solve the problem. Experimental results on a variety of both real and synthetic datasets confirm that our new robust algorithm considerably outperforms state-of-the-art metric learning methods in the presence of label noise and outliers.

关键词： Metric learning Rescaled hinge loss robust algorithm Label noise Outlier Half quadratic (HQ) optimization

来源：评论

学校读者我要写书评

暂无评论

Active Exploration in robust Unmanned Vehicle Task Assignment

引用

JOURNAL OF AEROSPACE COMPUTING INFORMATION AND COMMUNICATION 2011年第8期8卷 250-268页

作者： Bertuccelli, Luca F. How, Jonathan P. MIT Dept Aeronaut & Astronaut Cambridge MA 02139 USA

This paper presents new formulations for the robust allocation of unmanned vehicles (UVs) in the presence of parametric uncertainty. Standard robust optimization approaches for hedging against the worst-case plan performance can lead to overly conservative plans. One way to reduce this conservatism is to employ active exploration to reduce the uncertainty in the model parameters, but the UV literature does not address the coupling between uncertainty reduction and improvement in worst-case mission performance. This paper presents a new algorithm that assigns active exploration tasks that provide the most benefit in reducing the conservatism of the robust plans. This algorithm is extended to a more complex, receding horizon task assignment framework which has been validated in previous UV hardware demonstrations. We then show that this novel formulation can be formulated as an integer program and improves overall mission score when compared to alternative decoupled formulations. Finally, this paper shows that this coupled formulation demonstrates more tightly coordinated behavior between heterogeneous UVs in complex resource allocation problems.

关键词： Unmanned Vehicles Kalman Filter Monte Carlo Simulation Weapon Target Assignment Linear Programming Gamma Distribution Artificial Intelligence Sensitivity Analysis robust algorithm Adaptive Control

来源：评论

学校读者我要写书评

暂无评论

robust Sybil attack defense with information level in online Recommender Systems

引用

EXPERT SYSTEMS WITH APPLICATIONS 2014年第4期41卷 1781-1791页

作者： Noh, Giseop Kang, Young-myoung Oh, Hayoung Kim, Chong-kwon Seoul Natl Univ Dept Comp Sci & Engn Seoul 151744 South Korea Samsung Elect Suwon 443742 Gyeonggi Do South Korea Soongsil Univ Sch Elect & Engn Seoul 156743 South Korea

As the major function of Recommender Systems (RSs) is recommending commercial items to potential consumers (i.e., system users), providing correct information of RS is crucial to both RS providers and system users. The influence of RS over Online Social Networks (OSNs) is expanding rapidly, whereas malicious users continuously try to attack the RSs with fake identities (i.e.. Sybils) by manipulating the information in the RS adversely. In this paper, we propose a novel robust recommendation algorithm called RobuRec which exploits a distinctive feature, admission control. RobuRec provides highly trusted recommendation results since RobuRec predicts appropriate recommendations regardless of whether the ratings are given by honest users or by Sybils thanks to the power of admission control. To demonstrate the performance of RobuRec, we have conducted extensive experiments with various datasets as well as diverse attack scenarios. The evaluation results confirm that RobuRec outperforms the comparable schemes such as PCA and LTSMF significantly in terms of Prediction Shift (PS) and Hit Ratio (HR). (C) 2013 Elsevier Ltd. All rights reserved.

关键词： Sybil attack Recommendation systems robust algorithm

来源：评论

学校读者我要写书评

暂无评论

Toward Trustworthy Social Network Services: A robust Design of Recommender Systems

引用

JOURNAL OF COMMUNICATIONS AND NETWORKS 2015年第2期17卷 145-156页

作者： Noh, Giseop Oh, Hayoung Lee, Kyu-haeng Kim, Chong-kwon Seoul Natl Univ Dept Comp Sci & Engn Seoul South Korea Soongsil Univ Sch Elect & Engn Seoul South Korea

In recent years, electronic commerce and online social networks (OSNs) have experienced fast growth, and as a result, recommendation systems (RSs) have become extremely common. Accuracy and robustness are important performance indexes that characterize customized information or suggestions provided by RSs. However, nefarious users may be present, and they can distort information within the RSs by creating fake identities (Sybils). Although prior research has attempted to mitigate the negative impact of Sybils, the presence of these fake identities remains an unsolved problem. In this paper, we introduce a new weighted link analysis and influence level for RSs resistant to Sybil attacks. Our approach is validated through simulations of a broad range of attacks, and it is found to outperform other state-of-the-art recommendation methods in terms of both accuracy and robustness.

关键词： Link analysis recommender systems robust algorithm Sybil attack

来源：评论

学校读者我要写书评

暂无评论

Low-rank robust online distance/similarity learning based on the rescaled hinge loss

引用

APPLIED INTELLIGENCE 2023年第1期53卷 634-657页

作者： Zabihzadeh, Davood Tuama, Amar Karami-Mollaee, Ali Mousavirad, Seyed Jalaleddin Hakim Sabzevari Univ Dept Comp Engn Sabzevar Iran Ferdowsi Univ Mashhad Engn Fac Comp Engn Dept Mashhad Razavi Khorasan Iran Hakim Sabzevari Univ Elect & Comp Engn Fac Sabzevar Iran

An important challenge in metric learning is scalability to both size and dimension of input data. Online metric learning algorithms are proposed to address this challenge. Existing methods are commonly based on Passive/Aggressive (PA) approach. Hence, they can rapidly process large volumes of data with an adaptive learning rate. However, these algorithms are based on the Hinge loss and so are not robust against outliers and label noise. We address the challenges by formulating the online Distance/Similarity learning problem with the robust Rescaled Hinge loss function. The proposed model is rather general and can be applied to any PA-based online Distance/Similarity algorithm. To achieve scalability to data dimension, we propose low-rank online Distance/Similarity methods that learn a rectangular projection matrix instead of a full Mahalanobis matrix. The low-rank approaches not only reduce the computational cost but also keep the discrimination power of the learned metrics. Also, current online methods usually assume training triplets or pairwise constraints exist in advance. However, this assumption does not hold, and generating triplets using available batch sampling methods is both time and space consuming. We address this issue by developing an efficient, yet effective robust one-pass triplet construction algorithm. We conduct several experiments on datasets from various applications. The results confirm that the proposed methods significantly outperform state-of-the-art online metric learning methods in the presence of label noise and outliers by a large margin.

关键词： Metric learning Rescaled hinge loss robust algorithm Label noise Online distance similarity learning One pass triplet construction

来源：评论

学校读者我要写书评

暂无评论

robust distributed adaptation under arctangent and maximum correntropy criterion

引用

SIGNAL IMAGE AND VIDEO PROCESSING 2024年第11期18卷 8105-8118页

作者： Wang, Shengwei Xu, Yurong Ci, Caihong Xu, Tianci Cui, Shuohao Chen, Hongquan Northwest Normal Univ Coll Comp Sci & Engn Lanzhou 730000 Peoples R China Northwest Minzu Univ Coll Med Coll Lanzhou 730000 Peoples R China Northwest Normal Univ Coll Phys & Elect Engn Lanzhou 730000 Peoples R China

In this paper, the problem of robust distributed estimation in undirected networks is explored in depth. The Arctangent framework has garnered widespread adoption for adaptive estimation in prior research, with numerous adaptive estimation algorithms stemming from this foundation. However, the integration of the Arctangent framework with the maximum correntropy criterion, and its potential for enhanced estimation results, remains unexplored. In this paper, we have conducted a study in distributed networks and proposed a new robust distributed estimation algorithm using the Arctangent framework and the maximum correntropy criterion. Simulation experiments demonstrate that the proposed algorithm exhibits superior performance compared to the benchmark algorithm in both Gaussian and impulsive noise environments. Finally, a theoretical analysis of the proposed algorithm is conducted.

关键词： Distributed Estimation robust algorithm Adaptive Filtering Maximum correntropy criterion

来源：评论

学校读者我要写书评

暂无评论

robust APPROXIMATION SCHEMES FOR CUBE PACKING

引用

SIAM JOURNAL ON OPTIMIZATION 2013年第2期23卷 1310-1343页

作者： Epstein, Leah Levin, Asaf Univ Haifa Dept Math IL-31905 Haifa Israel Technion Israel Inst Technol Fac Ind Engn & Management IL-32000 Haifa Israel

Square bin packing, where square items are to be packed into a minimum number of unit size squares, and its multidimensional generalization, cube packing, where d-dimensional cubic items are to be packed into a minimum number of unit size cubes of the same dimension, are well-studied generalizations of the classical one-dimensional bin packing problem. These problems are known to have asymptotic polynomial time approximation schemes (APTAS). In this paper we design robust approximation schemes for these problems. At each step, a robust algorithm receives a single input item to be added to the packing. The new item must be packed and the packing can be modified, but the total volume of items which may migrate between bins, or change their positions inside bins, must be bounded by a constant factor times the volume of the new item. That is, the solution is created by slightly adjusting the solution upon the arrival of each new item. Previous approximation schemes partition items into three types according to size, and rely heavily on the knowledge of the complete input before the implementation of this partition and the creation of the solution, as the definition of the set of medium items is based on the entire specific input. We introduce new methods which allow us to adjust the partition dynamically, and to modify the packing accordingly. Using these methods, we develop a robust APTAS for every dimension d, that is, a robust polynomial time algorithm for d-dimensional cube packing, which maintains an asymptotically (1 + epsilon)-approximate solution throughout this process.

关键词： multidimensional packing migration factor approximation scheme robust algorithm

来源：评论

学校读者我要写书评

暂无评论

ROCS: robust One-Bit Compressed Sensing With Application to Direction of Arrival

引用

IEEE TRANSACTIONS ON SIGNAL PROCESSING 2024年 72卷 2407-2420页

作者： Li, Xiao-Peng Shi, Zhang-Lei Huang, Lei So, Anthony Man-Cho So, Hing Cheung Shenzhen Univ State Key Lab Radio Frequency Heterogeneous Integr Shenzhen 518060 Peoples R China China Univ Petr East China Coll Sci Qingdao 266580 Peoples R China Chinese Univ Hong Kong Dept Syst Engn & Engn Management Hong Kong Peoples R China City Univ Hong Kong Dept Elect Engn Hong Kong Peoples R China

One-bit compressed sensing (1-bit CS) inherits the merits of traditional CS and further reduces the cost and burden on the hardware device via employing the 1-bit analog-to-digital converter. When the measurements do not involve sign flips caused by additive noise, most contemporary algorithms can attain excellent signal restoration. However, their recovery performance might significantly degrade if there is even a small portion of sign flips. In order to increase the estimation accuracy in noisy scenarios, we devise a new signal model for 1-bit CS to attain robustness against sign flips. Then, we give a double-sparsity optimization formulation of the restoration problem. Subsequently, we combine proximal alternating minimization and projected gradient descent to tackle the problem. Different from existing robust methodologies, our approach, referred to as robust one-bit CS (ROCS), does not require the number of sign flips. Furthermore, we analyze the convergence behavior of ROCS and show that the objective value and variable sequences converge. Numerical results using synthetic data demonstrate that ROCS is superior to the competing methods in terms of reconstruction error in noisy environments. ROCS is also applied to direction-of-arrival estimation and outperforms state-of-the-art approaches.

关键词： Noise Quantization (signal) Noise measurement Vectors Signal processing algorithms Minimization Estimation robust algorithm one-bit compressed sensing direction-of-arrival estimation l(0)-norm optimization

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：