基于图连通支配集的子图匹配优化算法

在子图匹配过程中,随着图规模不断增长,匹配时间呈现指数爆炸的趋势.对此,提出一种基于图连通支配集的子图匹配优化算法VF-SMDS.根据贪心算法构建查询图的最小连通支配子图;通过代价模型计算最小连通支配子图节点的匹配代价,构建最优k查询节点匹配序列;通过支配节点的结构特征缩小查询节点搜索空间范围,在数据图中遍历到满足要求的节点,得到最终答案集.实验将VF-SMDS与GADDI、SPath、VF2++、VF3和SubISO方法进行对比.实验结果表明,在处理较大规模子图匹配问题时,VF-SMDS查询效率更高.     

使用社区结构信息的子图匹配算法优化方法*

子图匹配是图数据查询处理技术中的一个重要研究问题。针对现有子图匹配算法运行效率不高且缺乏通用优化方法的现状,提出一种基于社区结构的子图匹配算法优化方法(community structure based subgraph matching optimization method,CSO)。首先,提出两种优化策略,即解析模式图信息以减少子图匹配过程的计算量,以及利用社区结构信息在子图匹配过程中进行剪枝;然后,结合上述两种优化策略提出基于社区结构的子图匹配算法优化方法,并进行了理论分析。真实数据集和合成数据集上的大量实验结果表明,CSO方法能有效减少子图匹配算法的时间开销。同时,不同规模数据集上的实验结果验证了CSO方法良好的可扩展性。     

基于包含度的子图匹配方法

子图匹配是图论中最基本的操作.研究子图匹配的一个变种,即：在一个节点拥有若干元素的大图数据库中,找到与给定查询图结构同构并且对应节点元素的加权集合包含度大于给定值的所有子图,称作基于包含度的子图匹配（subgraph matching with inclusion degree,简称SMID）.该查询能够应用于多种场景,包括论文检索、社区发现、企业招聘等.为高效实现SMID,设计了同时包含节点元素和图结构信息的数据签名与查询签名,在离线处理阶段,利用数据签名为数据图建立动态签名树（DS-Tree）,以加快在线处理时图节点的匹配过程.为解决DS-Tree占用空间大的问题,设计了一种DS-Tree压缩方法,在对查询效率影响不大的情况下减小了索引空间.为进一步加快查询效率,还提出了支配子图查询算法.在真实数据和人工数据上的实验结果表明,所提出的方法在效率和扩展性方面优于现有其他方法.     

不确定图间α-β子图同构匹配算法

子图查询返回图数据集合中所有包含查询图的数据图.在查询图和数据图同时为不确定性图的前提下,提出了不确定图间的期望子图同构定义和α-β子图同构匹配定义.不确定图间的期望子图同构是确定图上子图同构在概率图模型上的直接推广,不确定图间α-β子图同构利用两个限制阈值来衡量查询图和数据图间的匹配质量.文章详细阐述了α-β子图同构匹配的语义特点,分析了其和期望子图同构的联系和差别,设计实现α-β子图同构匹配判定算法.     

计算机视觉中图匹配研究进展: 从二图匹配迈向多图匹配

图匹配试图求解二图或多图之间节点的对应关系.在图像图形领域,图匹配是一个历久弥新的基础性问题.从优化的角度来看,图匹配问题是一个组合优化问题,且在一般情形下具有非确定性多项式复杂程度(non-deter-ministic polynomial, NP)难度的性质.在过去数十年间,出现了大量求解二图匹配的近似算法,并在各个领域得到了较为广泛的应用.然而,受限于优化问题本身的理论困难和实际应用中数据质量的种种限制,各二图匹配算法在匹配精度上的性能日益趋近饱和.相比之下,由于引入了更多信息且往往更符合实际问题的设定,多图的协同匹配则逐渐成为了一个新兴且重要的研究方向.本文首先介绍了经典的二图匹配方法,随后着重介绍近年来多图匹配方法的最新进展和相关工作.最后,本文讨论了图匹配未来的发展.     

基于邻域等价类的同构子图搜索算法

节点异质图常作为复杂网络的数据模型,同构子图搜索是异质图挖掘过程中的重要问题,但现有算法的子图去重步骤降低了搜索效率。为此,基于Turbo_(ISO)算法中的邻域等价类(NEC)概念,提出同构子图搜索算法NEC-COMB。该算法包含预处理、节点顺序确定、子图同构匹配和子图提取4个部分,在子图同构匹配时对NEC中的节点使用组合策略,避免等价节点重复匹配。实验结果表明,与经典算法VF2,GraphQL,Turbo_(ISO)相比,NEC-COMB可有效提高搜索效率,优化去重效果。     

子图相似性的恶意程序检测方法

动态行为分析是一种常见的恶意程序分析方法,常用图来表示恶意程序系统调用或资源依赖等,通过图挖掘算法找出已知恶意程序样本中公共的恶意特征子图,并通过这些特征子图对恶意程序进行检测.然而这些方法往往依赖于图匹配算法,且图匹配不可避免计算慢,同时,算法中还忽视了子图之间的关系,而考虑子图间的关系有助于提高模型检测效果.为了解决这两个问题,提出了一种基于子图相似性恶意程序检测方法,即DMBSS.该方法使用数据流图来表示恶意程序运行时的系统行为或事件,再从数据流图中提取出恶意行为特征子图,并使用“逆拓扑标识”算法将特征子图表示成字符串,字符串蕴含了子图的结构信息,使用字符串替代图的匹配.然后,通过神经网络来计算子图间的相似性即将子图结构表示成高维向量,使得相似子图在向量空间的距离也较近.最后,使用子图向量构建恶意程序的相似性函数,并在此基础上,结合SVM分类器对恶意程序进行检测.实验结果显示,与其他方法相比,DMBSS在检测恶意程序时速度较快,且准确率较高.     

基于信息熵的子图匹配算法

子图查询是指输入一个图数据库和查询子图,输出图数据库中包含查询子图的图集合,它广泛应用于社会网、生物网和信息网的查询应用中。目前的子图查询算法大多采用静态消耗测算模式,此类测算模式在图中点数和连接边数呈指数分布时,会在少数节点上花费较多时间遍历其邻节点,导致查询算法效率低下。根据信息熵在信息度量中的作用,将条件信息熵作为启发式匹配的依据,提出了基于信息熵的子图匹配算法。实验表明,基于信息熵的子图匹配算法具有更高的查询效率,且在指数分布的数据集上效果更明显。     

基于持久性内存的属性图存储系统

属性图是一种流行的图数据模型, 在各种图系统中得到了广泛应用. 然而, 面向事务型负载的图数据库系统在执行图分析任务的场景下面临着高延迟等挑战. 传统的图分析系统往往是基于简单图模型, 而且大多不支持图的事务型负载. 因此, 迫切需要一个能够在属性图上高效处理事务型负载和图分析任务的图存储系统. 持久性内存的问世, 使得我们有机会重新设计图存储系统, 以充分发挥这种设备的特点. 为此, 本文提出了一种基于持久性内存的属性图存储系统, 名为TAG. TAG采用了一种新颖的混合架构的图存储方式, 以充分发挥持久性内存和主存的优势. 其次, 通过拓扑和索引结合的方式, 将图的拓扑嵌入到系统的索引中以加速图的拓扑查询. 最后, TAG通过基于标签的方式来组织图的属性数据, 进一步优化图的属性访问. 实验结果表明, TAG显著优于其他图数据库系统, 与图分析系统相比, TAG也有着相近的性能表现.     

面向子图匹配的社会网络隐私保护方法

云平台作为存储和处理分析大规模社会网络数据的工具逐渐变为主流,针对大规模社会网络子图匹配隐私保护问题,提出分布式K-自同构社会网络隐私保护算法,保护上传至云平台的社会网络图的结构隐私。通过节点间传递标记信息的方式添加噪声边,使原始图匿名为具有k个对称子图的K-自同构社会网络图。提出分布式的子图匹配方法对上传图进行子图匹配,根据搜索图中节点的选择性对搜索图进行分解得到搜索分解子图;在每个计算节点内进行分布并行的子图匹配得到搜索分解子图匹配结果,将结果连接后得到关于搜索图的匹配结果;在客户端中根据K-自同构社会网络图的对称性和K-自同构函数对得到的子图匹配结果进行恢复和过滤得到正确匹配结果。实验结果表明:分布式K-自同构社会网络隐私保护算法和分布式子图匹配方法在处理大规模社会网络图时具有很高的效率,并且有效解决了隐私泄露问题。     

Matching graphs with unique node labels

A special class of graphs is introduced in this paper. The graphs belonging to this class are characterised by the existence of unique node labels. A number of matching algorithms for graphs with unique node labels are developed. It is shown that problems such as graph isomorphism, subgraph isomorphism, maximum common subgraph (MCS) and graph edit distance (GED) have a computational complexity that is only quadratic in the number of nodes. Moreover, computing the median of a set of graphs is only linear in the cardinality of the set. In a series of experiments, it is demonstrated that the proposed algorithms run very fast in practice. The considered class makes the matching of large graphs, consisting of thousands of nodes, computationally tractable. We also discuss an application of the considered class of graphs and related matching algorithms to the classification and detection of abnormal events in computer networks.     

WSM: a novel algorithm for subgraph matching in large weighted graphs

Given an undirected/directed large weighted data graph and a similar smaller weighted pattern graph, the problem of weighted subgraph matching is to find a mapping of the nodes in the pattern graph to a subset of nodes in the data graph such that the sum of edge weight differences is minimum. Biological interaction networks such as protein-protein interaction networks and molecular pathways are often modeled as weighted graphs in order to account for the high false positive rate occurring intrinsically during the detection process of the interactions. Nonetheless, complex biological problems such as disease gene prioritization and conserved phylogenetic tree construction largely depend on the similarity calculation among the networks. Although several existing methods provide efficient methods for graph and subgraph similarity measurement, they produce nonintuitive results due to the underlying unweighted graph model assumption. Moreover, very few algorithms exist for weighted graph matching that are applicable with the restriction that the data and pattern graph sizes are equal. In this paper, we introduce a novel algorithm for weighted subgraph matching which can effectively be applied to directed/undirected weighted subgraph matching. Experimental results demonstrate the superiority and relative scalability of the algorithm over available state of the art methods.     

多编码树GPU并行轴心子图匹配

子图同构问题是非确定多项式(NP)完全问题,而轴心子图同构是一种特殊的子图同构问题.针对现在已经有许多高效的子图同构算法,然而对于轴心子图同构问题目前并没有基于GPU的搜索算法,且通过改造已有的子图同构算法来解决轴心子图匹配问题会产生大量不必要的中间结果这一问题,提出了一种基于GPU的轴心子图同构算法.首先,通过一种新...     

基于邻居信息聚合的子图同构匹配算法

图匹配在现实中被广泛运用,而子图同构匹配是其中的研究热点,具有重要的科学意义与实践价值。现有子图同构匹配算法大多基于邻居关系来构建约束条件,而忽略了节点的局部邻域信息。对此,提出了一种基于邻居信息聚合的子图同构匹配算法。首先,将图的属性和结构导入到改进的图卷积神经网络中进行特征向量的表示学习,从而得到聚合后的节点局部邻域信息;然后,根据图的标签、度等特征对匹配顺序进行优化,以提高算法的效率;最后,将得到的特征向量和优化的匹配顺序与搜索算法相结合,建立子图同构的约束满足问题（CSP）模型,并结合CSP回溯算法对模型进行求解。实验结果表明,与经典的树搜索算法和约束求解算法相比,该算法可以有效地提高子图同构的求解效率。     

Matching Feature Points in Image Sequences through a Region-Based Method

In order to achieve reliable matching in image sequences, a hierarchical approach is proposed. First, matching is established between regions of consecutive segmented images. In a second step, feature point matching between features of matched regions is performed. Regions provide a structural representation for features in the same region. To establish matching between features of two corresponding regions, a relational graph is built. Matching features is based on three principles: exclusion, proximity, and rigidity. The use of subgraph matching techniques through maximal clique detection provides a method to cope with partial occlusion and missing features between frames.     

Structural matching of 2D electrophoresis gels using deformed graphs

2D electrophoresis is a well-known method for protein separation which is extremely useful in the field of proteomics. Each spot in the image represents a protein accumulation and the goal is to perform a differential analysis between pairs of images to study changes in protein content. It is thus necessary to register two images by finding spot correspondences. Although it may seem a simple task, generally, the manual processing of this kind of images is very cumbersome, especially when strong variations between corresponding sets of spots are expected (e.g. strong non-linear deformations and outliers). In order to solve this problem, this paper proposes a new quadratic assignment formulation together with a correspondence estimation algorithm based on graph matching which takes into account the structural information between the detected spots. Each image is represented by a graph and the task is to find a maximum common subgraph. Successful experimental results using real data are presented, including an extensive comparative performance evaluation with ground-truth data.     

不确定图间α-β子图同构匹配算法

子图查询返回图数据集合中所有包含查询图的数据图。在查询图和数据图同时为不确定性图的前提下,提出了不确定图间的期望子图同构定义和α-β子图同构匹配定义。不确定图间的期望子图同构是确定图上子图同构在概率图模型上的直接推广,不确定图间α-β子图同构利用两个限制阈值来衡量查询图和数据图间的匹配质量。文章详细阐述了α-β子图同构匹配的语义特点,分析了其和期望子图同构的联系和差别,设计实现α-β子图同构匹配判定算法。     

MR-SimLab: Scalable subgraph selection with label similarity for big data

With the increasing size and complexity of available databases, existing machine learning and data mining algorithms are facing a scalability challenge. In many applications, the number of features describing the data could be extremely high. This hinders or even could make any further exploration infeasible. In fact, many of these features are redundant or simply irrelevant. Hence, feature selection plays a key role in helping to overcome the problem of information overload especially in big data applications. Since many complex datasets could be modeled by graphs of interconnected labeled elements, in this work, we are particularly interested in feature selection for subgraph patterns. In this paper, we propose MR-SimLab, a MapReduce-based approach for subgraph selection from large input subgraph sets. In many applications, it is easy to compute pairwise similarities between labels of the graph nodes. Our approach leverages such rich information to measure an approximate subgraph matching by aggregating the elementary label similarities between the matched nodes. Based on the aggregated similarity scores, our approach selects a small subset of informative representative subgraphs. We provide a distributed implementation of our algorithm on top of the MapReduce framework that optimizes the computational efficiency of our approach for big data applications. We experimentally evaluate MR-SimLab on real datasets. The obtained results show that our approach is scalable and that the selected subgraphs are informative.     

Hybrid query execution engine for large attributed graphs

Graphs are widely used for modeling complicated data such as social networks, bibliographical networks and knowledge bases. The growing sizes of graph databases motivate the crucial need for developing powerful and scalable graph-based query engines. We propose a SPARQL-like language, G-SPARQL, for querying attributed graphs. The language enables the expression of different types of graph queries that are of large interest in the databases that are modeled as large graph such as pattern matching, reachability and shortest path queries. Each query can combine both structural predicates and value-based predicates (on the attributes of the graph nodes/edges). We describe an algebraic compilation mechanism for our proposed query language which is extended from the relational algebra and based on the basic construct of building SPARQL queries, the Triple Pattern. We describe an efficient hybrid Memory/Disk representation of large attributed graphs where only the topology of the graph is maintained in memory while the data of the graph are stored in a relational database. The execution engine of our proposed query language splits parts of the query plan to be pushed inside the relational database (using SQL) while the execution of other parts of the query plan is processed using memory-based algorithms, as necessary. Experimental results on real and synthetic datasets demonstrate the efficiency and the scalability of our approach and show that our approach outperforms native graph databases by several factors.