ID 原文 译文
57078 本文方法为SAR图像目标三维重建提供了新的技术途径,是SAR目标三维重建的有益尝试. The proposed method is a useful attempt for 3D target reconstruction of SAR images and represents a new technical approach.
57079 群智感知系统通过对高维感知数据的发布和分析为人们带来巨大数据价值的同时,也给参与者的隐私带来了极大的隐患. High-dimensional crowdsourced data is pervasive in crowdsensing systems and with the developmentof IoTs it can produce rich knowledge for the society.
57080 目前,各种基于差分隐私的隐私保护方法被提出,但大部分方法不能同时解决高维感知数据间复杂的属性关联问题和来自不可信服务器的隐私威胁问题. However, it also creates serious privacy threats for crowd?sourcing participants. To mitigate the privacy concerns in crowdsensing systems, local differential privacy hasbeen derived from the de facto standard of differential privacy in order to achieve strong privacy guaranteed indistributed systems. However, directly achieving local differential privacy on high-dimensional crowdsourced datamay lead not only to a prohibitive computational burden but also low data utility.
57081 基于此,本文提出了基于Bayes网络的高维感知数据本地隐私保护发布机制.该机制实现了用户端的本地数据保护,杜绝了其他方直接访问用户原始数据的可能,根本上保护了用户的数据隐私.感知服务器端在接收到用户本地隐私保护的数据后,基于Bayes网络方法对高维数据的维度相关性进行识别,将高维数据属性集划分为多个相对独立的低维属性集,进而依次合成新的数据集,可以有效地保留原始感知数据的属性维度相关性,保证合成数据集与原始数据集具有尽可能相似的统计特性. Therefore, in this paper, wepropose a local private high-dimensional data publication scheme for crowdsensing systems. In particular, on theparticipants’ side, high-dimensional records are locally perturbed to protect privacy, while on the server’s side, theprobability distribution of original data is recovered by taking advantage of both the expectation maximizationalgorithm and the theory of the Bayesian network.
57082 通过大量仿真实验验证了该方法的有效性,实验结果表明该方法在有效的本地隐私保护下的合成数据具有较高的数据效用性. Extensive experiments on real-world datasets demonstratedthe effectiveness of the proposed scheme that can synthesize approximate datasets with local differential privacy.
57083 聚类是数据挖掘和机器学习中的基本任务之一. Clustering is one of fundamental tasks of data mining and machine learning.
57084 传统聚类方法由于其设计中对簇结构假设的限制,导致算法在不符合其假设的数据集上,尤其是大型高维数据集上的聚类效果较差. Due to the limitationof cluster assumption, lots of clustering algorithms perform poorly on some datasets against their assumptions,especially high-dimensional big data. This paper presents a maximum average entropy-rate based correlationclustering algorithm which is a kind of a graph-based correlation clustering.
57085 本文引入了最大平均熵率的概念,设计了一种基于图的关联聚类算法. The objective function of originalcorrelation clustering is decomposed into several single cluster optimizations and the limitation of big data incorrelation clustering is removed by the neighboring connected graph.
57086 该算法将关联聚类问题分解为多个独立的单类优化问题,并利用邻域消除了关联聚类对大数据的限制. In algorithm implementation, the opti?mization of proper neighbor searching and correlation clustering are performed by heuristic neighbor searchingand cluster generating respectively, and there is also an efficient graph-iterated implementation on distributedcomputation platform.
57087 算法实现通过启发式邻域搜索和类生成简化了对最优邻域和关联聚类的求解过程,并且设计了适应分布式计算平台的图迭代方法. Compared with other clustering algorithms, the proposed clustering algorithm is moreflexible in cluster assumption, when accelerating the clustering process.