针对监督学习方法采集攻击样本困难以及无监督学习方法检测精度不足的问题,提出一种融合自监督学习与主动学习的域名系统(domain name system,DNS)隧道检测方法。该方法采用异常检测框架,无需获取攻击样本,同时,通过自监督学习引入训练...针对监督学习方法采集攻击样本困难以及无监督学习方法检测精度不足的问题,提出一种融合自监督学习与主动学习的域名系统(domain name system,DNS)隧道检测方法。该方法采用异常检测框架,无需获取攻击样本,同时,通过自监督学习引入训练指导过程,通过主动学习引入反馈调节过程,显著提升了检测精度。构建基于Transformer架构的自编码器,通过对正常样本特征进行自监督学习,实现了DNS数据包级别的异常检测。以此为基础,将主动学习方法应用于反馈引导的孤立森林(feedback-guided isolated forest,FBIF),实现了DNS交互流级别的异常检测,将检出的异常流视为与隧道攻击活动相关。实验结果表明,该检测方法在无需获取攻击样本的前提下,能准确检测出多种类型的隧道攻击,且在资源消耗方面具备高可扩展性。展开更多
域名系统(domain name system,DNS)隐蔽信道是一种利用DNS协议实现数据泄露的网络攻击手段,受到诸多高级持续性威胁(advanced persistent threat,APT)组织的青睐,给网络空间安全带来了严重威胁。针对传统机器学习方法对特征依赖性强、...域名系统(domain name system,DNS)隐蔽信道是一种利用DNS协议实现数据泄露的网络攻击手段,受到诸多高级持续性威胁(advanced persistent threat,APT)组织的青睐,给网络空间安全带来了严重威胁。针对传统机器学习方法对特征依赖性强、误报率高的问题,提出一种融合多通道卷积和注意力网络的DNS隐蔽信道检测算法。该算法基于DNS请求与响应双向流,首先将残差结构和并行卷积相结合,采用不同大小的卷积核提取并融合多尺度特征信息,实现不同感受野特征的捕获;其次引入通道注意力机制增加卷积通道关键信息的提取能力,丰富网络模型的表达能力;最后采用softmax函数实现DNS隐蔽信道的检测。实验结果表明,所提模型能有效检测DNS隐蔽信道,平均准确率、精确率和召回率分别为96.42%、97.82%和96.16%,优于传统方法。展开更多
In this paper,the Paley-Wiener theorem is extended to the analytic function spaces with general weights.We first generalize the theorem to weighted Hardy spaces Hp(0<p<∞)on tube domains by constructing a sequen...In this paper,the Paley-Wiener theorem is extended to the analytic function spaces with general weights.We first generalize the theorem to weighted Hardy spaces Hp(0<p<∞)on tube domains by constructing a sequence of L^(1)functions converging to the given function and verifying their representation in the form of Fourier transform to establish the desired result of the given function.Applying this main result,we further generalize the Paley-Wiener theorem for band-limited functions to the analytic function spaces L^(p)(0<p<∞)with general weights.展开更多
Pb(Zr,Ti)O_(3)-Pb(Zn_(1/3)Nb_(2/3))O_(3) (PZT-PZN) based ceramics, as important piezoelectric materials, have a wide range of applications in fields such as sensors and actuators, thus the optimization of their piezoe...Pb(Zr,Ti)O_(3)-Pb(Zn_(1/3)Nb_(2/3))O_(3) (PZT-PZN) based ceramics, as important piezoelectric materials, have a wide range of applications in fields such as sensors and actuators, thus the optimization of their piezoelectric properties has been a hot research topic. This study investigated the effects of phase boundary engineering and domain engineering on (1-x)[0.8Pb(Zr_(0.5)Ti_(0.5))O_(3)-0.2Pb(Zn_(1/3)Nb_(2/3))O_(3)]-xBi(Zn_(0.5)Ti_(0.5))O_(3) ((1-x)(0.8PZT-0.2PZN)- xBZT) ceramic to obtain excellent piezoelectric properties. The crystal phase structure and microstructure of ceramic samples were characterized. The results showed that all samples had a pure perovskite structure, and the addition of BZT gradually increased the grain size. The addition of BZT caused a phase transition in ceramic samples from the morphotropic phase boundary (MPB) towards the tetragonal phase region, which is crucial for optimizing piezoelectric properties. By adjusting content of BZT and precisely controlling position of the phase boundary, the piezoelectric performance can be optimized. Domain structure is one of the key factors affecting piezoelectric performance. By using domain engineering techniques to optimize grain size and domain size, piezoelectric properties of ceramic samples have been significantly improved. Specifically, excellent piezoelectric properties (piezoelectric constant d_(33)=320 pC/N, electromechanical coupling factor kp=0.44) were obtained simultaneously for x=0.08. Based on experimental results and theoretical analysis, influence mechanisms of phase boundary engineering and domain engineering on piezoelectric properties were explored. The study shows that addition of BZT not only promotes grain growth, but also optimizes the domain structure, enabling the polarization reversal process easier, thereby improving piezoelectric properties. These research results not only provide new ideas for the design of high-performance piezoelectric ceramics, but also lay a theoretical foundation for development of related electronic devices.展开更多
Automatically recognizing radar emitters from com-plex electromagnetic environments is important but non-trivial.Moreover,the changing electromagnetic environment results in inconsistent signal distribution in the rea...Automatically recognizing radar emitters from com-plex electromagnetic environments is important but non-trivial.Moreover,the changing electromagnetic environment results in inconsistent signal distribution in the real world,which makes the existing approaches perform poorly for recognition tasks in different scenes.In this paper,we propose a domain generaliza-tion framework is proposed to improve the adaptability of radar emitter signal recognition in changing environments.Specifically,we propose an end-to-end denoising based domain-invariant radar emitter recognition network(DDIRNet)consisting of a denoising model and a domain invariant representation learning model(IRLM),which mutually benefit from each other.For the signal denoising model,a loss function is proposed to match the feature of the radar signals and guarantee the effectiveness of the model.For the domain invariant representation learning model,contrastive learning is introduced to learn the cross-domain feature by aligning the source and unseen domain distri-bution.Moreover,we design a data augmentation method that improves the diversity of signal data for training.Extensive experiments on classification have shown that DDIRNet achieves up to 6.4%improvement compared with the state-of-the-art radar emitter recognition methods.The proposed method pro-vides a promising direction to solve the radar emitter signal recognition problem.展开更多
The ammunition loading system manipulator is susceptible to gear failure due to high-frequency,heavyload reciprocating motions and the absence of protective gear components.After a fault occurs,the distribution of fau...The ammunition loading system manipulator is susceptible to gear failure due to high-frequency,heavyload reciprocating motions and the absence of protective gear components.After a fault occurs,the distribution of fault characteristics under different loads is markedly inconsistent,and data is hard to label,which makes it difficult for the traditional diagnosis method based on single-condition training to generalize to different conditions.To address these issues,the paper proposes a novel transfer discriminant neural network(TDNN)for gear fault diagnosis.Specifically,an optimized joint distribution adaptive mechanism(OJDA)is designed to solve the distribution alignment problem between two domains.To improve the classification effect within the domain and the feature recognition capability for a few labeled data,metric learning is introduced to distinguish features from different fault categories.In addition,TDNN adopts a new pseudo-label training strategy to achieve label replacement by comparing the maximum probability of the pseudo-label with the test result.The proposed TDNN is verified in the experimental data set of the artillery manipulator device,and the diagnosis can achieve 99.5%,significantly outperforming other traditional adaptation methods.展开更多
面对ISP主干网,为了检测威胁其管理域内用户安全的僵尸网络、钓鱼网站以及垃圾邮件等恶意活动,实时监测流经主干网边界的DNS交互报文,并从域名的依赖性和使用位置两个方面刻画DNS活动行为模式,而后,基于有监督的多分类器模型,提出面向IS...面对ISP主干网,为了检测威胁其管理域内用户安全的僵尸网络、钓鱼网站以及垃圾邮件等恶意活动,实时监测流经主干网边界的DNS交互报文,并从域名的依赖性和使用位置两个方面刻画DNS活动行为模式,而后,基于有监督的多分类器模型,提出面向ISP主干网的上层DNS活动监测算法DAOS(binary classifier for DNS activity observation system).其中,依赖性从用户角度观察域名的外在使用情况,而使用位置则关注区域文件中记录的域名内部资源配置.实验结果表明:该算法在不依赖先验知识的前提下,经过两小时的DNS活动观测,可以达到90.5%的检测准确率,以及2.9%的假阳性和6.6%的假阴性.若持续观察1周,准确率可以上升到93.9%,假阳性和假阴性也可以下降到1.3%和4.8%.展开更多
文摘针对监督学习方法采集攻击样本困难以及无监督学习方法检测精度不足的问题,提出一种融合自监督学习与主动学习的域名系统(domain name system,DNS)隧道检测方法。该方法采用异常检测框架,无需获取攻击样本,同时,通过自监督学习引入训练指导过程,通过主动学习引入反馈调节过程,显著提升了检测精度。构建基于Transformer架构的自编码器,通过对正常样本特征进行自监督学习,实现了DNS数据包级别的异常检测。以此为基础,将主动学习方法应用于反馈引导的孤立森林(feedback-guided isolated forest,FBIF),实现了DNS交互流级别的异常检测,将检出的异常流视为与隧道攻击活动相关。实验结果表明,该检测方法在无需获取攻击样本的前提下,能准确检测出多种类型的隧道攻击,且在资源消耗方面具备高可扩展性。
文摘域名系统(domain name system,DNS)隐蔽信道是一种利用DNS协议实现数据泄露的网络攻击手段,受到诸多高级持续性威胁(advanced persistent threat,APT)组织的青睐,给网络空间安全带来了严重威胁。针对传统机器学习方法对特征依赖性强、误报率高的问题,提出一种融合多通道卷积和注意力网络的DNS隐蔽信道检测算法。该算法基于DNS请求与响应双向流,首先将残差结构和并行卷积相结合,采用不同大小的卷积核提取并融合多尺度特征信息,实现不同感受野特征的捕获;其次引入通道注意力机制增加卷积通道关键信息的提取能力,丰富网络模型的表达能力;最后采用softmax函数实现DNS隐蔽信道的检测。实验结果表明,所提模型能有效检测DNS隐蔽信道,平均准确率、精确率和召回率分别为96.42%、97.82%和96.16%,优于传统方法。
基金Supported by the National Natural Science Foundation of China(12301101)the Guangdong Basic and Applied Basic Research Foundation(2022A1515110019 and 2020A1515110585)。
文摘In this paper,the Paley-Wiener theorem is extended to the analytic function spaces with general weights.We first generalize the theorem to weighted Hardy spaces Hp(0<p<∞)on tube domains by constructing a sequence of L^(1)functions converging to the given function and verifying their representation in the form of Fourier transform to establish the desired result of the given function.Applying this main result,we further generalize the Paley-Wiener theorem for band-limited functions to the analytic function spaces L^(p)(0<p<∞)with general weights.
基金National Natural Science Foundation of China (52202139, 52072178)。
文摘Pb(Zr,Ti)O_(3)-Pb(Zn_(1/3)Nb_(2/3))O_(3) (PZT-PZN) based ceramics, as important piezoelectric materials, have a wide range of applications in fields such as sensors and actuators, thus the optimization of their piezoelectric properties has been a hot research topic. This study investigated the effects of phase boundary engineering and domain engineering on (1-x)[0.8Pb(Zr_(0.5)Ti_(0.5))O_(3)-0.2Pb(Zn_(1/3)Nb_(2/3))O_(3)]-xBi(Zn_(0.5)Ti_(0.5))O_(3) ((1-x)(0.8PZT-0.2PZN)- xBZT) ceramic to obtain excellent piezoelectric properties. The crystal phase structure and microstructure of ceramic samples were characterized. The results showed that all samples had a pure perovskite structure, and the addition of BZT gradually increased the grain size. The addition of BZT caused a phase transition in ceramic samples from the morphotropic phase boundary (MPB) towards the tetragonal phase region, which is crucial for optimizing piezoelectric properties. By adjusting content of BZT and precisely controlling position of the phase boundary, the piezoelectric performance can be optimized. Domain structure is one of the key factors affecting piezoelectric performance. By using domain engineering techniques to optimize grain size and domain size, piezoelectric properties of ceramic samples have been significantly improved. Specifically, excellent piezoelectric properties (piezoelectric constant d_(33)=320 pC/N, electromechanical coupling factor kp=0.44) were obtained simultaneously for x=0.08. Based on experimental results and theoretical analysis, influence mechanisms of phase boundary engineering and domain engineering on piezoelectric properties were explored. The study shows that addition of BZT not only promotes grain growth, but also optimizes the domain structure, enabling the polarization reversal process easier, thereby improving piezoelectric properties. These research results not only provide new ideas for the design of high-performance piezoelectric ceramics, but also lay a theoretical foundation for development of related electronic devices.
基金supported by the National Natural Science Foundation of China(62101575)the Research Project of NUDT(ZK22-57)the Self-directed Project of State Key Laboratory of High Performance Computing(202101-16).
文摘Automatically recognizing radar emitters from com-plex electromagnetic environments is important but non-trivial.Moreover,the changing electromagnetic environment results in inconsistent signal distribution in the real world,which makes the existing approaches perform poorly for recognition tasks in different scenes.In this paper,we propose a domain generaliza-tion framework is proposed to improve the adaptability of radar emitter signal recognition in changing environments.Specifically,we propose an end-to-end denoising based domain-invariant radar emitter recognition network(DDIRNet)consisting of a denoising model and a domain invariant representation learning model(IRLM),which mutually benefit from each other.For the signal denoising model,a loss function is proposed to match the feature of the radar signals and guarantee the effectiveness of the model.For the domain invariant representation learning model,contrastive learning is introduced to learn the cross-domain feature by aligning the source and unseen domain distri-bution.Moreover,we design a data augmentation method that improves the diversity of signal data for training.Extensive experiments on classification have shown that DDIRNet achieves up to 6.4%improvement compared with the state-of-the-art radar emitter recognition methods.The proposed method pro-vides a promising direction to solve the radar emitter signal recognition problem.
文摘The ammunition loading system manipulator is susceptible to gear failure due to high-frequency,heavyload reciprocating motions and the absence of protective gear components.After a fault occurs,the distribution of fault characteristics under different loads is markedly inconsistent,and data is hard to label,which makes it difficult for the traditional diagnosis method based on single-condition training to generalize to different conditions.To address these issues,the paper proposes a novel transfer discriminant neural network(TDNN)for gear fault diagnosis.Specifically,an optimized joint distribution adaptive mechanism(OJDA)is designed to solve the distribution alignment problem between two domains.To improve the classification effect within the domain and the feature recognition capability for a few labeled data,metric learning is introduced to distinguish features from different fault categories.In addition,TDNN adopts a new pseudo-label training strategy to achieve label replacement by comparing the maximum probability of the pseudo-label with the test result.The proposed TDNN is verified in the experimental data set of the artillery manipulator device,and the diagnosis can achieve 99.5%,significantly outperforming other traditional adaptation methods.
文摘面对ISP主干网,为了检测威胁其管理域内用户安全的僵尸网络、钓鱼网站以及垃圾邮件等恶意活动,实时监测流经主干网边界的DNS交互报文,并从域名的依赖性和使用位置两个方面刻画DNS活动行为模式,而后,基于有监督的多分类器模型,提出面向ISP主干网的上层DNS活动监测算法DAOS(binary classifier for DNS activity observation system).其中,依赖性从用户角度观察域名的外在使用情况,而使用位置则关注区域文件中记录的域名内部资源配置.实验结果表明:该算法在不依赖先验知识的前提下,经过两小时的DNS活动观测,可以达到90.5%的检测准确率,以及2.9%的假阳性和6.6%的假阴性.若持续观察1周,准确率可以上升到93.9%,假阳性和假阴性也可以下降到1.3%和4.8%.
