L1投影的解析计算方法

点到平面距离的解析表示对度量间隔、模式可分性起到决定性作用,该距离均可归结为范数最小化问题.除L2范数易于求解外,其他类型范数求解均困难.以L1范数为例,尽管L1范数问题是凸的,由于L1范数的不可导性,迄今尚无解析表示,所以目前的L1学习机并非从L1间隔导出.讨论了在L1赋范线性空间中,L1距离及在超平面上的投影解析计算问题,主要完成了:(1)导出了L1范数下的点到超平面距离以及点在平面上的投影的解析表达式;(2)证明了该投影与欧氏度量下的L2范数投影之间的关系,并给出了几何解释.最后通过模拟实验,验证解析解的正确性及计算效率.

基于组合故障频繁树的最小失效诱因模式定位方法

针对定位实际软件中最小失效诱因模式可能受到屏蔽效应影响的问题,提出了一种基于组合故障频繁树的最小失效诱因模式定位方法及其迭代框架.该方法首先依据组合测试用例集及测试结果构建组合故障频繁树,然后从组合故障频繁树中抽取频繁参数值组合作为可疑失效诱因模式,并根据其可疑得分进行排序.基于给出的失效诱因模式迭代定位框架,反复迭代直到满足某一个停止准则为止.利用仿真实验对存在和不存在掩蔽效应影响的2种情形进行有效性验证.实验结果表明,在这2种情形下所提方法均能定位最小失效诱因模式,有效减少附加测试用例的数目.

基于故障森林的组合测试故障定位研究

组合测试作为一种对参数组合空间抽样的系统方法,适用于待测系统中存在由特定参数组合所引发的软件失效.依据组合测试结果,定位出最小失效诱因模式(minimal failure-causing schema,MFS)有助于程序员进行故障源检测与修复.然而,组合测试可能存在mask effect,使得测试用例中即使包含MFS也未必一定触发软件失效.因此,在存在mask effect的系统中精确定位最小失效诱因模式尤为困难.为此提出了一种基于故障森林的组合测试故障定位方法.给定一个t-路组合测试集(t≥2)及其附加测试集,该方法首先学习由多个深度为t的基本故障分类树所组成的故障森林,然后从故障森林中提取基本故障组合模式,最后将可疑MFS进行排序,并提交给程序员进行进一步诊断.仿真实验结果表明,该方法能有效定位系统中存在的组合故障模式.特别地,对于存在mask effect的待测系统,故障定位结果健壮.

改进的基于二次型模糊c均值聚类模型

针对模糊聚类算法对点数据集聚类的敏感性以及区间类型数据聚类效果不明显等问题,提出了基于二次型距离改进的模糊可能性c均值(fuzzy-possibilistic c-means,FPCM)聚类算法。首先分析了区间数据的特征,引入了区间值的数学表示方法,在此基础上提出了3种不同的基于区间数据距离度量方法以及相应权重矩阵的计算方法,通过建立拉格朗日方程对目标方程优化,求得聚类中心、隶属度以及可能性迭代方程,并证明目标方程的收敛性,最后给出了算法执行步骤。在不同类型的数据集上实验,证明算法在点数据集和区间数据集上都具有较好聚类性能.

一种非噪声敏感性的模糊C均值聚类算法

FCM算法提出用模糊隶属度表示样本数据的隶属于某个类的程度,能够克服HCM算法划分的不合理性.研究分析发现FCM算法对噪声数据具有敏感性,很难有效的识别噪声数据;FCM由于限制条件使得聚类结果与实际的分类不一致.针对此不足之处,文章提出一种非噪声敏感性FCM算法(INFCM),取消了限制条件,用典型值代替了隶属度值,构建了目标方程,为了克服聚类过程中一致性,在目标方程中增加了惩罚因子,分析了惩罚因子的组成,最后提出了聚类算法步骤.实验表明新的聚类算法能够有效克服对噪声的敏感性,提高了聚类的可理解性.

Identifying Similar Operation Scenes for Busy Area Sector Dynamic Management

Air traffic controllers face challenging initiatives due to uncertainty in air traffic.One way to support their initiatives is to identify similar operation scenes.Based on the operation characteristics of typical busy area control airspace,an complexity measurement indicator system is established.We find that operation in area sector is characterized by aggregation and continuity,and that dimensionality and information redundancy reduction are feasible for dynamic operation data base on principle components.Using principle components,discrete features and time series features are constructed.Based on Gaussian kernel function,Euclidean distance and dynamic time warping(DTW)are used to measure the similarity of the features.Then the matrices of similarity are input in Spectral Clustering.The clustering results show that similar scenes of trend are not ideal and similar scenes of modes are good base on the indicator system.Finally,actual vertical operation decisions for area sector and results of identification are compared,which are visualized by metric multidimensional scaling(MDS)plots.We find that identification results can well reflect the operation at peak hours,but controllers make different decisions under the similar conditions before dawn.The compliance rate of busy operation mode and division decisions at peak hours is 96.7%.The results also show subjectivity of actual operation and objectivity of identification.In most scenes,we observe that similar air traffic activities provide regularity for initiatives,validating the potential of this approach for initiatives and other artificial intelligence support.

Recognition of Similar Weather Scenarios in Terminal Area Based on Contrastive Learning

In order to improve the recognition accuracy of similar weather scenarios(SWSs)in terminal area,a recognition model for SWS based on contrastive learning(SWS-CL)is proposed.Firstly,a data augmentation method is designed to improve the number and quality of weather scenarios samples according to the characteristics of convective weather images.Secondly,in the pre-trained recognition model of SWS-CL,a loss function is formulated to minimize the distance between the anchor and positive samples,and maximize the distance between the anchor and the negative samples in the latent space.Finally,the pre-trained SWS-CL model is fine-tuned with labeled samples to improve the recognition accuracy of SWS.The comparative experiments on the weather images of Guangzhou terminal area show that the proposed data augmentation method can effectively improve the quality of weather image dataset,and the proposed SWS-CL model can achieve satisfactory recognition accuracy.It is also verified that the fine-tuned SWS-CL model has obvious advantages in datasets with sparse labels.

Identification of Similar Air Traffic Scenes with Active Metric Learning

The rapid growth of air traffic has continuously increased the workload of controllers,which has become an important factor restricting sector capacity.If similar traffic scenes can be identified,the historical decision-making experience may be used to help controllers decide control strategies quickly.Considering that there are many traffic scenes and it is hard to label them all,in this paper,we propose an active SVM metric learning(ASVM2L)algorithm to measure and identify the similar traffic scenes.First of all,we obtain some traffic scene samples correctly labeled by experienced air traffic controllers.We design an active sampling strategy based on voting difference to choose the most valuable unlabeled samples and label them.Then the metric matrix of all the labeled samples is learned and used to complete the classification of traffic scenes.We verify the effectiveness of ASVM2L on standard data sets,and then use it to measure and classify the traffic scenes on the historical air traffic data set of the Central South Sector of China.The experimental results show that,compared with other existing methods,the proposed method can use the information of traffic scene samples more thoroughly and achieve better classification performance under limited labeled samples.

Handling Label Noise in Air Traffic Complexity Evaluation Based on Confident Learning and XGBoost

Air traffic complexity is a critical indicator for air traffic operation,and plays an important role in air traffic management(ATM),such as airspace reconfiguration,air traffic flow management and allocation of air traffic controllers(ATCos).Recently,many machine learning techniques have been used to evaluate air traffic complexity by constructing a mapping from complexity related factors to air traffic complexity labels.However,the low quality of complexity labels,which is named as label noise,has often been neglected and caused unsatisfactory performance in air traffic complexity evaluation.This paper aims at label noise in air traffic complexity samples,and proposes a confident learning and XGBoost-based approach to evaluate air traffic complexity under label noise.The confident learning process is applied to filter out noisy samples with various label probability distributions,and XGBoost is used to train a robust and high-performance air traffic complexity evaluation model on the different label noise filtered ratio datasets.Experiments are carried out on a real dataset from the Guangzhou airspace sector in China,and the results prove that the appropriate label noise removal strategy and XGBoost algorithm can effectively mitigate the label noise problem and achieve better performance in air traffic complexity evaluation.

A Hybrid Method of Extractive Text Summarization Based on Deep Learning and Graph Ranking Algorithms

In the era of Big Data,we are faced with an inevitable and challenging problem of“overload information”.To alleviate this problem,it is important to use effective automatic text summarization techniques to obtain the key information quickly and efficiently from the huge amount of text.In this paper,we propose a hybrid method of extractive text summarization based on deep learning and graph ranking algorithms(ETSDG).In this method,a pre-trained deep learning model is designed to yield useful sentence embeddings.Given the association between sentences in raw documents,a traditional LexRank algorithm with fine-tuning is adopted fin ETSDG.In order to improve the performance of the extractive text summarization method,we further integrate the traditional LexRank algorithm with deep learning.Testing results on the data set DUC2004 show that ETSDG has better performance in ROUGE metrics compared with certain benchmark methods.

An Intelligent Early Warning Method of Press-Assembly Quality Based on Outlier Data Detection and Linear Regression

Focusing on controlling the press-assembly quality of high-precision servo mechanism,an intelligent early warning method based on outlier data detection and linear regression is proposed.Linear regression is used to deal with the relationship between assembly quality and press-assembly process,then the mathematical model of displacement-force in press-assembly process is established and a qualified press-assembly force range is defined for assembly quality control.To preprocess the raw dataset of displacement-force in the press-assembly process,an improved local outlier factor based on area density and P weight(LAOPW)is designed to eliminate the outliers which will result in inaccuracy of the mathematical model.A weighted distance based on information entropy is used to measure distance,and the reachable distance is replaced with P weight.Experiments show that the detection efficiency of the algorithm is improved by 5.6 ms compared with the traditional local outlier factor(LOF)algorithm,and the detection accuracy is improved by about 2%compared with the local outlier factor based on area density(LAOF)algorithm.The application of LAOPW algorithm and the linear regression model shows that it can effectively carry out intelligent early warning of press-assembly quality of high precision servo mechanism.

Image Deraining for UAV Using Split Attention Based Recursive Network

Images captured in rainy days suffer from noticeable degradation of scene visibility.Unmanned aerial vehicles(UAVs),as important outdoor image acquisition systems,demand a proper rain removal algorithm to improve visual perception quality of captured images as well as the performance of many subsequent computer vision applications.To deal with rain streaks of different sizes and directions,this paper proposes to employ convolutional kernels of different sizes in a multi-path structure.Split attention is leveraged to enable communication across multiscale paths at feature level,which allows adaptive receptive field to tackle complex situations.We incorporate the multi-path convolution and the split attention operation into the basic residual block without increasing the channels of feature maps.Moreover,every block in our network is unfolded four times to compress the network volume without sacrificing the deraining performance.The performance on various benchmark datasets demonstrates that our method outperforms state-of-the-art deraining algorithms in both numerical and qualitative comparisons.

Evolutionary Algorithm with Ensemble Classifier Surrogate Model for Expensive Multiobjective Optimization

For many real-world multiobjective optimization problems,the evaluations of the objective functions are computationally expensive.Such problems are usually called expensive multiobjective optimization problems(EMOPs).One type of feasible approaches for EMOPs is to introduce the computationally efficient surrogates for reducing the number of function evaluations.Inspired from ensemble learning,this paper proposes a multiobjective evolutionary algorithm with an ensemble classifier(MOEA-EC)for EMOPs.More specifically,multiple decision tree models are used as an ensemble classifier for the pre-selection,which is be more helpful for further reducing the function evaluations of the solutions than using single inaccurate model.The extensive experimental studies have been conducted to verify the efficiency of MOEA-EC by comparing it with several advanced multiobjective expensive optimization algorithms.The experimental results show that MOEA-EC outperforms the compared algorithms.

Double Transformed Tubal Nuclear Norm Minimization for Tensor Completion

Non-convex methods play a critical role in low-rank tensor completion for their approximation to tensor rank is tighter than that of convex methods.But they usually cost much more time for calculating singular values of large tensors.In this paper,we propose a double transformed tubal nuclear norm(DTTNN)to replace the rank norm penalty in low rank tensor completion(LRTC)tasks.DTTNN turns the original non-convex penalty of a large tensor into two convex penalties of much smaller tensors,and it is shown to be an equivalent transformation.Therefore,DTTNN could take advantage of non-convex envelopes while saving time.Experimental results on color image and video inpainting tasks verify the effectiveness of DTTNN compared with state-of-the-art methods.

基于访问均衡性核度中心的分布式密集无线网络边缘缓存决策策略

随着5G密集基站的部署,回程网络的能耗和延迟大大影响了用户体验。移动边缘网络缓存技术是解决回程负载、减少网络延迟、提高用户体验的有效途径。然而,边缘缓存网络中内容缓存位置的选择对网络缓存效率有很大的影响。现有算法没有考虑基站节点间的协作及动态的访问变化,因此在满足访问效率方面表现不够理想。基于此,提出一种基于访问均衡性核度中心的缓存决策方案,以提高对用户请求的响应。仿真结果表明,与现有机制相比,该方案在满足用户请求方面具有更好的性能。

Transition Stage Control of Tail Sitter Aircraft Based on Guardian Maps

To deal with the high nonlinearities and strong couplings in the transition stage of tailsitter aircraft,an adaptive gainscheduling controller is proposed by combining the guardian maps theory and H∞control theory.This method is applied to track the flightpath angle of the transition stage of tailsitter aircraft,and compared with the linear quadratic regulator(LQR)method based on traditional gain scheduling.Simulation results show that the controller based on the guardian maps theory can autonomously schedule the appropriate control parameters and accomplish the stable transition.Besides,the proposed method shows better tracking performance than the LQR method based on traditional gain scheduling.

编译优化序列选择研究进展

在过去的几十年里,编译器开发者针对各种复杂情况下的编译优化需求,设计实现了大量的编译优化选项.在实际开发中,由编译器提供的标准编译优化序列难以适应复杂场景下待编译程序的编译要求.一方面,待编译程序有不同的语义和编译目标,直接采用标准编译优化序列难以获得理想的优化效果,若采用不适当的优化序列甚至可能对程序性能等带来负面影响.另一方面,随着硬件体系结构的不断发展,编译环境日益复杂,编译优化序列亦应进行相应调整.因此,如何在错综复杂的优化选项中为待编译程序选择最佳的编译优化序列成为一个具有挑战性的科学问题.针对上述问题,研究人员展开了大量的研究,并取得了诸多成果.本文旨在归纳编译优化序列选择领域的研究文献,通过文献搜索,筛选获得符合条件的55篇论文,从多个视角(算法、研究类型、目标编译器、基准测试集等)揭示该领域的研究现状.通过文献分析可以发现,当前该领域的主流算法包括两类,即以遗传算法为代表的启发式搜索算法和以支持向量机为代表的机器学习算法.超过80%的文献的研究类型属于提出解决方案或者实证研究.在已有的研究中,实验验证时使用频次最多的编译器和基准测试集分别是GCC和miBench.本文有助于理解编译优化序列选择领域当前基本进展和发展趋势,同时为开展该领域研究工作提供了可能的方向.