检索结果-南通市图书馆

Towards the universal defense for query-based audio adversarial attacks on speech recognition system

Cybersecurity 2024年第1期7卷 53-70页

作者： Feng Guo Zheng Sun Yuxuan Chen Lei Ju School of Cyber Science and Technology Shandong UniversityQingdaoChina Quancheng Laboratory QCLJinanChina

Recently,studies show that deep learning-based automatic speech recognition(ASR)systems are vulnerable to adversarial examples(AEs),which add a small amount of noise to the original audio examples.These AE attacks pose new challenges to deep learning security and have raised significant concerns about deploying ASR systems and devices.The existing defense methods are either limited in application or only defend on results,but not on process.In this work,we propose a novel method to infer the adversary intent and discover audio adversarial examples based on the AEs generation process.The insight of this method is based on the observation:many existing audio AE attacks utilize query-based methods,which means the adversary must send continuous and similar queries to target ASR models during the audio AE generation process.Inspired by this observation,We propose a memory mechanism by adopting audio fingerprint technology to analyze the similarity of the current query with a certain length of memory query.Thus,we can identify when a sequence of queries appears to be suspectable to generate audio AEs.Through extensive evaluation on four state-of-the-art audio AE attacks,we demonstrate that on average our defense identify the adversary’s intent with over 90%accuracy.With careful regard for robustness evaluations,we also analyze our proposed defense and its strength to withstand two adaptive attacks.Finally,our scheme is available out-of-the-box and directly compatible with any ensemble of ASR defense models to uncover audio AE attacks effectively without model retraining.

关键词： Adversarial attacks Defense memory mechanism Query-based

来源：

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Hybrid artificial bee colony algorithm with variable neighborhood search and memory mechanism

引用

Journal of Systems Engineering and Electronics 2018年第2期29卷 405-414页

作者： FAN Chengli FU Qiang LONG Guangzheng XING Qinghua Air and Missile Defense College Air Force Engineering University

Artificial bee colony(ABC) is one of the most popular swarm intelligence optimization algorithms which have been widely used in numerical optimization and engineering applications. However, there are still deficiencies in ABC regarding its local search ability and global search efficiency. Aiming at these deficiencies,an ABC variant named hybrid ABC(HABC) algorithm is proposed.Firstly, the variable neighborhood search factor is added to the solution search equation, which can enhance the local search ability and increase the population diversity. Secondly, inspired by the neuroscience investigation of real honeybees, the memory mechanism is put forward, which assumes the artificial bees can remember their past successful experiences and further guide the subsequent foraging behavior. The proposed memory mechanism is used to improve the global search efficiency. Finally, the results of comparison on a set of ten benchmark functions demonstrate the superiority of HABC.

关键词： artificial bee colony(ABC) hybrid artificial bee colony(HABC) variable neighborhood search factor memory mechanism

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

A membrane-inspired algorithm with a memory mechanism for knapsack problems

引用

Journal of Zhejiang University-Science C(Computers and Electronics) 2013年第8期14卷 612-622页

作者： Juan-juan HE Jian-hua XIAO Xiao-long SHI Tao SONG Key Laboratory of Image Processing and Intelligent Control School of AutomationHuazhong University of Science and Technology The Logistics Research Center Nankai University

Membrane algorithms are a class of distributed and parallel algorithms inspired by the structure and behavior of living cells. Many attractive features of living cells have already been abstracted as operators to improve the performance of algorithms. In this work, inspired by the function of biological neuron cells storing information, we consider a memory mechanism by introducing memory modules into a membrane algorithm. The framework of the algorithm consists of two kinds of modules (computation modules and memory modules), both of which are arranged in a ring neighborhood topology. They can store and process information, and exchange information with each other. We test our method on a knapsack problem to demonstrate its feasibility and effectiveness. During the process of approaching the optimum solution, feasible solutions are evolved by rewriting rules in each module, and the information transfers according to directions defined by communication rules. Simulation results showed that the performance of membrane algorithms with memory cells is superior to that of algorithms without memory cells for solving a knapsack problem. Furthermore, the memory mechanism can prevent premature convergence and increase the possibility of finding a global solution.

关键词： Membrane algorithm memory mechanism Knapsack problem

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

A Hysteresis Modeling Approach Using “Shape Function + memory mechanism” Method

A Hysteresis Modeling Approach Using “Shape Function + Memo...

引用

第十四届全国流变学学术会议

作者：白先旭陈朋合肥工业大学

Hysteresis is an intrinsic property of magnetorheological（MR） fluid actuators（see figure 1）, no matter whether it works in shear mode, or flow mode, or squeeze mode, and it restrains the wide applications of MR technology when a precise hysteresis model is absent. The experimental researches and analysis show that, especially from the view of absolute continuity of input and rate-independence of input-output curves, MR fluid actuators behave as a kind of specific Duhem hysteresis where a memory effect is involved. Correspondently, several Duhem hysteresis models, like Bouc-Wen, Dahl and LuGre models, and their modified ones are applied to describe the behavior of MR fluids actuators. However, in these models the differential expression makes it trivial to solve the equations and the specific format of the model limits the flexibility to model the variable hysteresis behavior of MR fluid actuators. In this paper, a new hysteresis model based on shape function and memory mechanism is proposed and its feasibility is verified through modeling the hysteresis behavior of a commercial MR damper. Before the presentation of the new hysteresis model, the memory behavior and rate-independence of the hysteresis in MR damper and a new hysteresis phenomenon（see figure 2）, in resistor-capacitor（RC） circuit are presented. In the proposed model, the memory mechanism originating from the charging and discharging processes of the RC circuit is constructed by adopting a virtual displacement variable and updating laws for the reference points. The shape function is achieved and generalized from analytical solutions of the simple semi-linear Duhem model. Using the proposed approach, the memory mechanism reveals the essence of specific Duhem model and the general shape function provides a direct and clear means to fit the hysteresis loop. Eventually, within the structure of a "Restructured model"（see figure 3）, the original hysteresis model, Bouc-Wen model, is replaced by the proposed one, and with a predefined the shape function, the

关键词： Hysteresis modeling memory mechanism shape function RC circuit computational efficiency

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

请选择保存的检索档案：

请选择收藏分类：

通借通还

欢迎您,

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

在线全文

在线全文

在线全文

在线全文

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：