On web user tracking of browsing patterns for personalised advertising

On today’s Web, users trade access to their private data for content and services. App and service providers want to know everything they can about their users, in order to improve their product experience. Also, advertising sustains the business model of many websites and applications. Efficient and successful advertising relies on predicting users’ actions and tastes to suggest a range of products to buy. Both service providers and advertisers try to track users’ behaviour across their product network. For application providers this means tracking users’ actions within their platform. For third-party services following users, means being able to track them across different websites and applications. It is well known how, while surfing the Web, users leave traces regarding their identity in the form of activity patterns and unstructured data. These data constitute what is called the user’s online footprint. We analyse how advertising networks build and collect users footprints and how the suggested advertising reacts to changes in the user behaviour.     

隐私保护的数据发布研究

随着信息技术的发展,个人隐私泄露成为日益严重的问题,因此迫切需要研究防止数据发布中个人隐私的泄露。为此,许多研究者提出不同的方法用以实现隐私保护的数据发布。为总结前人工作,介绍了隐私保护数据发布技术的研究意义和发展历程,阐述了本领域研究过程中的背景攻击模型和隐私模型,深入分析了用已有的概化/隐匿方法和聚类方法实现匿名数据发布技术,总结了匿名质量有关的信息度量标准,同时探讨了数据更新引起的增量数据发布方法和高维数据、移动数据的发布,最后归纳了目前研究中的问题并展望了本领域进一步的研究趋势。     

Answering range-based reverse kNN and why-not reverse kNN queries

1 Introduction Given a dataset D and a point q,a reverse nearest neighbor(RNN)query retrieves all the points p∈D that have q as their nearest neighbor.Although the RNN problem was first pro-posed in[1],it still has received considerable attention due to its importance in several applications involving decision support,resource allocation,profile-based marketing,etc.     

基于缺陷度量的软件质量成本分析模型

针对软件项目的质量成本效益无法直接度量的这一难题,提出了建立组织级质量成本度量基线的解决方案。设计了基于缺陷度量的软件质量成本分析模型,实现了对软件质量成本及其效益的定量度量与分析,为评价软件质量保证活动绩效、指导质量投资决策提供了试用方法。     

Privacy preserving publication of relational and transaction data: Survey on the anonymization of patient data

Past two decades have seen a growing interest in methods for providing privacy preserving data publishing. Several models, algorithms and system designs have been proposed in the literature to protect identities in the published data. A number of these have been implemented and are successfully in use by diverse applications such as medical, supermarkets and e-commerce. This work presents a comprehensive survey of the previous research done to develop techniques for ensuring privacy while publishing data. We have identified and described in detail the concepts, models and algorithms related to this problem. The emphasis of the present work, in particular, is upon preserving privacy of patient data that includes demographics data, diagnosis codes and the data containing both demographics and diagnosis codes. Finally, we have summarized some of the major open problems in privacy preserving data publishing and discussed the possible directions for future work in this domain.     

Coverage metrics for temporal logic model checking*

In formal verification, we verify that a system is correct with respect to a specification. Even when the system is proved to be correct, there is still a question of how complete the specification is, and whether it really covers all the behaviors of the system. In this paper we study coverage metrics for model checking. Coverage metrics are based on modifications we apply to the system in order to check which parts of it were actually relevant for the verification process to succeed. We introduce two principles that we believe should be part of any coverage metric for model checking: a distinction between state-based and logic-based coverage, and a distinction between the system and its environment. We suggest several coverage metrics that apply these principles, and we describe two algorithms for finding the non-covered parts of the system under these definitions. The first algorithm is a symbolic implementation of a naive algorithm that model checks many variants of the original system. The second algorithm improves the naive algorithm by exploiting overlaps in the variants. We also suggest a few helpful outputs to the user, once the non-covered parts are found.

Coverage metrics for formal verification

In formal verification, we verify that a system is correct with respect to a specification. Even when the system is proven to be correct, there is still a question of how complete the specification is and whether it really covers all the behaviors of the system. The challenge of making the verification process as exhaustive as possible is even more crucial in simulation-based verification, where the infeasible task of checking all input sequences is replaced by checking a test suite consisting of a finite subset of them. It is very important to measure the exhaustiveness of the test suite, and indeed there has been extensive research in the simulation-based verification community on coverage metrics, which provide such a measure. It turns out that no single measure can be absolute, leading to the development of numerous coverage metrics whose usage is determined by industrial verification methodologies. On the other hand, prior research of coverage in formal verification has focused solely on state-based coverage. In this paper we adapt the work done on coverage in simulation-based verification to the formal-verification setting in order to obtain new coverage metrics. Thus, for each of the metrics used in simulation-based verification, we present a corresponding metric that is suitable for the setting of formal verification and describe an algorithmic way to check it.     

On some questions regarding k -regular and k -context-free sequences

We answer two questions of Allouche and Shallit regarding k-regular sequences and k-context-free sequences. One of these has been solved independently in another way by Bell. We also provide a partial solution to another question of Allouche and Shallit regarding the subword complexity of k-context-free sequences.     

Ray representation for k-trees

k-trees have established themselves as useful data structures in pattern recognition. A fundamental operation regarding k-trees is the construction of a k-tree. We present a method to store an object as a set of rays and an algorithm to convert such a set into a k-tree. The algorithm is conceptually simple, works for any k, and builds a k-tree from the rays very fast. It produces a minimal k-tree and does not lead to intermediate storage swell.     

On the k-ary hypercube

In this paper, we propose and analyze a new interconnection network, the k-ary hypercube. This new architecture captures the advantages of the mesh network and those of the binary hypercube. We show that the hamiltoniacity of this network and its capability of efficiently simulating other topologies. It has a smaller degree than that of its equivalent binary hypercube (the one with at least as many nodes) and has a smaller diameter than its equivalent mesh of processors.     

支持隐私保护的k近邻分类器

k近邻（k-nearest neighbor,简称kNN）分类器在生物信息学、股票预测、网页分类以及鸢尾花分类预测等方面都有着广泛的应用.随着用户隐私保护意识的日益提高,kNN分类器也需要对密文数据提供分类支持,进而保证用户数据的隐私性,即设计一种支持隐私保护的k近邻分类器（privacy-preserving k-nearest neighbor classifier,简称PP-kNN）.首先,对kNN分类器的操作进行分析,从中提取出一些基本操作,包括加法、乘法、比较、内积等.然后,选择两种同态加密方案和一种全同态加密方案对数据进行加密.在此基础上设计了针对基本操作的安全协议,其输出结果与在明文数据上执行同一方法的输出结果一致,且证明该协议在半诚实模型下是安全的.最后,通过将基本操作的安全协议进行模块化顺序组合的方式实现kNN分类器对密文数据处理的支持.通过实验,对所设计的PP-kNN分类器进行测试.结果表明,该分类器能够以较高效率实现对密文数据的分类,同时为用户数据提供隐私性保护.     

Adaptive soft k-nearest-neighbour classifiers

A novel classifier is introduced to overcome the limitations of the k-NN classification systems. It estimates the posterior class probabilities using a local Parzen window estimation with the k-nearest-neighbour prototypes (in the Euclidean sense) to the pattern to classify. A learning algorithm is also presented to reduce the number of data points to store. Experimental results in two hand-written classification problems demonstrate the potential of the proposed classification system.     

The accelerated k-in-a-row game

We investigate two versions of the well-known k-in-a-row game. While in the most intriguing k = 5 case the outcome of the game has been recently settled, very little is known about what happens when the rules are changed. A natural modification is that the players take more than one square of the board per move in order to speed up the game. Our main goal is to improve the quadratic bound on the error term, given by Csirmaz in Csirmaz (Discrete Math. 29 (1980) 19–23), to a logarithmic one for the accelerated k-in-a-row. The other issue is the extreme sensitivity of k-in-a-row under biased rules. Beck proposed in Beck (unpublished lecture notes) that a player may trade some of his freedom of choice for the right of taking more squares than his opponent. We prove logarithmic bounds on the error term in that case, too.     

基于k-Lin假设的同态加密方案

作为数字货币的底层核心技术之一,区块链随着数字货币的快速发展而受到了广泛关注.由于区块链具有去中心化、防篡改、可追溯等性质,如今越来越多的企业和个人用户选择利用区块链技术来实现数据的传输和记录.区块链公开透明的特性,一方面充分保证了数据的可用性;但另一方面,又给用户的隐私信息带来了严重威胁.为了同时兼顾用户数据的机密性和可用性,同态加密常常被用到区块链的安全解决方案之中.然而,现实应用对于所部署的同态加密方案的安全强度要求也很可能会随着时间推移而有所变化.考虑到区块链应用场景的复杂多样性和分布式特点,同态加密方案一旦部署下去,之后,当随着时间推移需要调整安全性强度时,相应的工作量将会非常繁重.此外,在区块链的现实应用中,考虑到监管方面的需求,很多情况下（尤其是针对某些群组成员发布和传输的数据）需要允许某可信第三方（如监管方）能够对链上的相应密文数据进行解密.若采用传统的同态加密方案对数据进行加密,可信第三方需要存储所有用户的私钥,这将给密钥管理和存储带来巨大压力.针对当前的区块链应用场景和安全需求,提出了一个基于ZN2*（N=pq）上的判定性k-Lin假设的加法同态加密方案.该方案不仅在标准模型下能够满足IND-CCA1安全性,还具有3个特殊优势：（i）可以通过对参数k的调控细粒度地调节加密方案的安全性强度;（ii）加密方案具有双解密机制：存在两种私钥,一种由用户本人持有,另一种由可信第三方持有,其中,可信第三方的私钥可用于该加密体制所有用户的密文解密;（iii）加密方案可以极为便利地退化为IND-CPA安全的公钥加密方案,退化后的方案不仅其公私钥长度和密文长度变得更短,而且同样具有加法同态性和双解密机制.     

应用k-means算法实现标记分布学习

标记分布学习是近年来提出的一种新的机器学习范式,它能很好地解决某些标记多义性的问题。现有的标记分布学习算法均利用条件概率建立参数模型,但未能充分利用特征和标记间的联系。本文考虑到特征相似的样本所对应的标记分布也应当相似,利用原型聚类的k均值算法（k-means）,将训练集的样本进行聚类,提出基于k-means算法的标记分布学习（label distribution learning based on k-means algorithm,LDLKM）。首先通过聚类算法k-means求得每一个簇的均值向量,然后分别求得对应标记分布的均值向量。最后将测试集和训练集的均值向量间的距离作为权重,应用到对测试集标记分布的预测上。在6个公开的数据集上进行实验,并与3种已有的标记分布学习算法在5种评价指标上进行比较,实验结果表明提出的LDLKM算法是有效的。     

面向云数据的隐私度量研究进展

隐私保护技术是云计算环境中防止隐私信息泄露的重要保障,通过度量这种泄露风险可反映隐私保护技术的隐私保护强度,以便构建更好的隐私保护方案。因此,隐私度量对隐私保护具有重大意义。主要对现有面向云数据的隐私度量方法进行综述：首先,对隐私保护技术和隐私度量进行概述,给出攻击者背景知识的量化方法,提出云数据隐私保护技术的性能评价指标和一种综合评估框架;然后,提出一种云数据隐私度量抽象模型,从工作原理和具体实施的角度对基于匿名、信息熵、集对分析理论和差分隐私四类隐私度量方法进行详细阐述;再从隐私度量指标和度量效果方面分析与总结这四类方法的优缺点及其适用范围;最后,从隐私度量的过程、效果和方法三方面指出云数据隐私度量技术的发展趋势及有待解决的问题。     

基于RBF神经网络的最经济控制研究

为有效解决系统的最经济控制问题,本文提出将系统的经济收益问题转换为对系统控制结构和参数的优化问题。首先提出将网络代价的概念植入径向基函数神经网络(RBF网络)结构的优化中,对神经网络的隐层激活函数和隐层节点数进行选择;再用改进的遗传算法实现RBF网络参数的优化,从而实现神经网络的最经济控制;最后通过实例验证,表明设计的算法与BP网络的最经济控制相对比,具有明显的优越性。     

Embedding k(n−k) edge-disjoint spanning trees in arrangement graphs

The arrangement graphs are a class of generalized star graphs. In this paper we construct a graph that consists of the maximum number of directed edge-disjoint spanning trees in an arrangement graph. The paths that connect the common root node to any given node through different spanning trees are node-disjoint, and the lengths of these paths differ from the shortest possible lengths by a small additive constant. This graph can be used to derive fault-tolerant algorithms for broadcasting and scattering problems without prior knowledge of the faulty elements of the network.     

Separating k-separated eNCE graph languages

An eNCE graph grammar is k-separated (k1) if the distance between any two nonterminal nodes in any of its sentential forms is at least k. Let SEP_k denote the class of graph languages generated by k-separated grammars. Then, SEP₁ (SEP₂) is the class of eNCE (boundary eNCE) graph languages, and so SEP₂SEP₁. Recently, Engelfriet (1991) showed that SEP₃SEP₂ and conjectured that, in fact, SEP_k+1SEP_k for each k 1. We prove this conjecture affirmatively.