Attributed Network Embedding with Micro-Meso Structure

Recently, network embedding has received a large amount of attention in network analysis. Although some network embedding methods have been developed from different perspectives, on one hand, most of the existing methods only focus on leveraging the plain network structure, ignoring the abundant attribute information of nodes. On the other hand, for some methods integrating the attribute information, only the lower-order proximities (e.g., microscopic proximity structure) are taken into account, which may suffer if there exists the sparsity issue and the attribute information is noisy. To overcome this problem, the attribute information and mesoscopic community structure are utilized. In this article, we propose a novel network embedding method termed Attributed Network Embedding with Micro-Meso structure, which is capable of preserving both the attribute information and the structural information including the microscopic proximity structure and mesoscopic community structure. In particular, both the microscopic proximity structure and node attributes are factorized by Nonnegative Matrix Factorization (NMF), from which the low-dimensional node representations can be obtained. For the mesoscopic community structure, a community membership strength matrix is inferred by a generative model (i.e., BigCLAM) or modularity from the linkage structure, which is then factorized by NMF to obtain the low-dimensional node representations. The three components are jointly correlated by the low-dimensional node representations, from which two objective functions (i.e., ANEM_B and ANEM_M) can be defined. Two efficient alternating optimization schemes are proposed to solve the optimization problems. Extensive experiments have been conducted to confirm the superior performance of the proposed models over the state-of-the-art network embedding methods.

A Graph Representation Learning Algorithm for Low-Order Proximity Feature Extraction to Enhance Unsupervised IDS Preprocessing

Most existing studies on an unsupervised intrusion detection system (IDS) preprocessing ignore the relationship among packets. According to the homophily hypothesis, the local proximity structure in the similarity relational graph has similar embedding after preprocessing. To improve the performance of IDS by building a relationship among packets, we propose a packet2vec learning algorithm that extracts accurate local proximity features based on graph representation by adding penalty to node2vec. In this algorithm, we construct a relational graph G’ by using each packet as a node, calculate the cosine similarity between packets as edges, and then explore the low-order proximity of each packet via the penalty-based random walk in G’. We use the above algorithm as a preprocessing method to enhance the accuracy of unsupervised IDS by retaining the local proximity features of packets maximally. The original features of the packet are combined with the local proximity features as the input of a deep auto-encoder for IDS. Experiments based on ISCX2012 show that the proposal outperforms the state-of-the-art algorithms by 11.6% with respect to the accuracy of unsupervised IDS. It is the first time to introduce graph representation learning for packet-embedded preprocessing in the field of IDS.

AUTOMATIC CORRECTION ALGORITHM OF HYFROLOGY FEATURE ATTRIBUTE IN NATIONAL GEOGRAPHIC CENSUS

A subset of the attributes of hydrologic features data in national geographic census are not clear, the current solution to this problem was through manual filling which is inefficient and liable to mistakes. So this paper proposes an automatic correction algorithm of hydrologic features attribute. Based on the analysis of the structure characteristics and topological relation, we put forward three basic principles of correction which include network proximity, structure robustness and topology ductility. Based on the WJ-III map workstation, we realize the automatic correction of hydrologic features. Finally, practical data is used to validate the method. The results show that our method is highly reasonable and efficient.

Host associations and turnover of haemosporidian parasites in manakins (Aves: Pipridae)

SUMMARYParasites of the genera Plasmodium and Haemoproteus (Apicomplexa: Haemosporida) are a diverse group of pathogens that infect birds nearly worldwide. Despite their ubiquity, the ecological and evolutionary factors that shape the diversity and distribution of these protozoan parasites among avian communities and geographic regions are poorly understood. Based on a survey throughout the Neotropics of the haemosporidian parasites infecting manakins (Pipridae), a family of Passerine birds endemic to this region, we asked whether host relatedness, ecological similarity and geographic proximity structure parasite turnover between manakin species and local manakin assemblages. We used molecular methods to screen 1343 individuals of 30 manakin species for the presence of parasites. We found no significant correlations between manakin parasite lineage turnover and both manakin species turnover and geographic distance. Climate differences, species turnover in the larger bird community and parasite lineage turnover in non-manakin hosts did not correlate with manakin parasite lineage turnover. We also found no evidence that manakin parasite lineage turnover among host species correlates with range overlap and genetic divergence among hosts. Our analyses indicate that host switching (turnover among host species) and dispersal (turnover among locations) of haemosporidian parasites in manakins are not constrained at this scale.

Population Structure

Population structure is a ubiquitous feature of natural populations that has an important influence on evolutionary change. In the real world, populations are not homogenous units; instead, they develop an internal structure, created by the physical properties of the environment and the biological characteristics of the species (such as dispersal ability). However, our basic ecological and population genetic models generally ignore population structure and focus on randomly mating (panmictic) populations. Such structure can profoundly change the evolution of a population. In fact, the myriad of influences that population structure exerts can only be hinted at in a single chapter. Since an exhaustive review is not possible, I will focus on presenting the conceptual issues linking mathematical models of population structure to empirical studies. To do this, it is useful to recognize two different kinds of population structure that both reflect and influence evolutionary change. The first is genetic structure. This is defined as the nonrandom distribution of genotypes in space and time. Thus, genetic structure reflects the genetic differences that develop among the different components of one or more populations. The second is what I will call proximity structure, defined by the size and composition of the group of neighbors that influence an individual’s fitness. Fitness is commonly influenced by local intraspecific interactions. Perhaps the most obvious example is competition. When individuals compete for some resource, they don’t usually compete equally with every other member of the population; in general, they compete only with a few of the most proximate individuals. These two forms of population structure, genetic structure and proximity structure, provide a foundation for understanding why we have shifted away from viewing populations as homogenous units. For good reason, this is a theme that is explored in many of the other chapters in this book. Genetic structure can develop within a population over a single generation, generally either as a result of local family associations or as a result of spatial variation in selection. For example, limited seed dispersal results in genetic correlations among neighbors even in the face of long-distance pollen movement, due to the clustering of maternal half sibs.