arXiv subjects

Marc Barthelemy

Publications and source records attributed to Marc Barthelemy.

At least 127 records · Page 7Linked to original sources

arXiv2005

Knowledge Representation Issues in Semantic Graphs for Relationship Detection

An important task for Homeland Security is the prediction of threat vulnerabilities, such as through the detection of relationships between seemingly disjoint entities. A structure used for this task is a "semantic graph", also known as a "relational data graph" or an "attributed relational graph". These graphs encode relationships as "typed" links between a pair of "typed" nodes. Indeed, semantic graphs are very similar to semantic networks used in AI. The node and link types are related through an ontology graph (also known as a schema). Furthermore, each node has a set of attributes associated with it (e.g., "age" may be an attribute of a node of type "person"). Unfortunately, the selection of types and attributes for both nodes and links depends on human expertise and is somewhat subjective and even arbitrary. This subjectiveness introduces biases into any algorithm that operates on semantic graphs. Here, we raise some knowledge representation issues for semantic graphs and provide some possible solutions using recently developed ideas in the field of complex networks. In particular, we use the concept of transitivity to evaluate the relevance of individual links in the semantic graph for detecting relationships. We also propose new statistical measures for semantic graphs and illustrate these semantic measures on graphs constructed from movies and terrorism data.

cs.AI↗

arXiv2005

The effects of spatial constraints on the evolution of weighted complex networks

Motivated by the empirical analysis of the air transportation system, we define a network model that includes geographical attributes along with topological and weight (traffic) properties. The introduction of geographical attributes is made by constraining the network in real space. Interestingly, the inclusion of geometrical features induces non-trivial correlations between the weights, the connectivity pattern and the actual spatial distances of vertices. The model also recovers the emergence of anomalous fluctuations in the betweenness-degree correlation function as first observed by Guimerà and Amaral [Eur. Phys. J. B {\bf 38}, 381 (2004)]. The presented results suggest that the interplay between weight dynamics and spatial constraints is a key ingredient in order to understand the formation of real-world weighted networks.

physics.soc-ph↗

arXiv2004

Dynamical patterns of epidemic outbreaks in complex heterogeneous networks

We present a thorough inspection of the dynamical behavior of epidemic phenomena in populations with complex and heterogeneous connectivity patterns. We show that the growth of the epidemic prevalence is virtually instantaneous in all networks characterized by diverging degree fluctuations, independently of the structure of the connectivity correlation functions characterizing the population network. By means of analytical and numerical results, we show that the outbreak time evolution follows a precise hierarchical dynamics. Once reached the most highly connected hubs, the infection pervades the network in a progressive cascade across smaller degree classes. Finally, we show the influence of the initial conditions and the relevance of statistical results in single case studies concerning heterogeneous networks. The emerging theoretical framework appears of general interest in view of the recently observed abundance of natural networks with complex topological features and might provide useful insights for the development of adaptive strategies aimed at epidemic containment.

cond-mat.dis-nn↗

arXiv2004

Characterization and Modeling of weighted networks

We review the main tools which allow for the statistical characterization of weighted networks. We then present two case studies, the airline connection network and the scientific collaboration network, which are representative of critical infrastructures and social systems, respectively. The main empirical results are (i) the broad distributions of various quantities and (ii) the existence of weight-topology correlations. These measurements show that weights are relevant and that in general the modeling of complex networks must go beyond topology. We review a model which provides an explanation for the features observed in several real-world networks. This model of weighted network formation relies on the dynamical coupling between topology and weights, considering the rearrangement of weights when new links are introduced in the system.

cond-mat.stat-mech↗

arXiv2004

Modeling the evolution of weighted networks

We present a general model for the growth of weighted networks in which the structural growth is coupled with the edges' weight dynamical evolution. The model is based on a simple weight-driven dynamics and a weights' reinforcement mechanism coupled to the local network growth. That coupling can be generalized in order to include the effect of additional randomness and non-linearities which can be present in real-world networks. The model generates weighted graphs exhibiting the statistical properties observed in several real-world systems. In particular, the model yields a non-trivial time evolution of vertices properties and scale-free behavior with exponents depending on the microscopic parameters characterizing the coupling rules. Very interestingly, the generated graphs spontaneously achieve a complex hierarchical architecture characterized by clustering and connectivity correlations varying as a function of the vertices' degree.

cond-mat.stat-mech↗

arXiv2004

Weighted evolving networks: coupling topology and weights dynamics

We propose a model for the growth of weighted networks that couples the establishment of new edges and vertices and the weights' dynamical evolution. The model is based on a simple weight-driven dynamics and generates networks exhibiting the statistical properties observed in several real-world systems. In particular, the model yields a non-trivial time evolution of vertices' properties and scale-free behavior for the weight, strength and degree distributions.

cond-mat.dis-nn↗

arXiv2004

Traffic-driven model of the World Wide Web graph

We propose a model for the World Wide Web graph that couples the topological growth with the traffic's dynamical evolution. The model is based on a simple traffic-driven dynamics and generates weighted directed graphs exhibiting the statistical properties observed in the Web. In particular, the model yields a non-trivial time evolution of vertices and heavy-tail distributions for the topological and traffic properties. The generated graphs exhibit a complex architecture with a hierarchy of cohesiveness levels similar to those observed in the analysis of real data.

cs.NI↗

arXiv2004

Betweenness Centrality in Large Complex Networks

We analyze the betweenness centrality (BC) of nodes in large complex networks. In general, the BC is increasing with connectivity as a power law with an exponent $η$. We find that for trees or networks with a small loop density $η=2$ while a larger density of loops leads to $η<2$. For scale-free networks characterized by an exponent $γ$ which describes the connectivity distribution decay, the BC is also distributed according to a power law with a non universal exponent $δ$. We show that this exponent $δ$ must satisfy the exact bound $δ\geq (γ+1)/2$. If the scale free network is a tree, then we have the equality $δ=(γ+1)/2$.

cond-mat.dis-nn↗

arXiv2004

Velocity and hierarchical spread of epidemic outbreaks in scale-free networks

We study the effect of the connectivity pattern of complex networks on the propagation dynamics of epidemics. The growth time scale of outbreaks is inversely proportional to the network degree fluctuations, signaling that epidemics spread almost instantaneously in networks with scale-free degree distributions. This feature is associated with an epidemic propagation that follows a precise hierarchical dynamics. Once the highly connected hubs are reached, the infection pervades the network in a progressive cascade across smaller degree classes. The present results are relevant for the development of adaptive containment strategies.

cond-mat.dis-nn↗

arXiv2003

The architecture of complex weighted networks

Networked structures arise in a wide array of different contexts such as technological and transportation infrastructures, social phenomena, and biological systems. These highly interconnected systems have recently been the focus of a great deal of attention that has uncovered and characterized their topological complexity. Along with a complex topological structure, real networks display a large heterogeneity in the capacity and intensity of the connections. These features, however, have mainly not been considered in past studies where links are usually represented as binary states, i.e. either present or absent. Here, we study the scientific collaboration network and the world-wide air-transportation network, which are representative examples of social and large infrastructure systems, respectively. In both cases it is possible to assign to each edge of the graph a weight proportional to the intensity or capacity of the connections among the various elements of the network. We define new appropriate metrics combining weighted and topological observables that enable us to characterize the complex statistical properties and heterogeneity of the actual strength of edges and vertices. This information allows us to investigate for the first time the correlations among weighted quantities and the underlying topological structure of the network. These results provide a better description of the hierarchies and organizational principles at the basis of the architecture of weighted networks.

cond-mat.stat-mech↗

arXiv2003

Comment on: 'Universal Behavior of Load Distribution in Scale-free Networks'

In a previous Letter (cond-mat/0106565), Goh et al have presented a numerical study of the load--or betweenness centrality--distribution in a scale-free network whose degree distribution follows a power law with a tunable exponent $γ$. They showed that the load $\ell$ is distributed according to a power-law with exponent $δ$. The authors claimed that $δ$ is universal, ie. independent of the exponent $γ$. In this comment, we use two different ways of checking numerically this universality and we show that it does not hold.

cond-mat.dis-nn↗

arXiv2003

Crossover from Scale-Free to Spatial Networks

In many networks such as transportation or communication networks, distance is certainly a relevant parameter. In addition, real-world examples suggest that when long-range links are existing, they usually connect to hubs-the well connected nodes. We analyze a simple model which combine both these ingredients--preferential attachment and distance selection characterized by a typical finite `interaction range'. We study the crossover from the scale-free to the `spatial' network as the interaction range decreases and we propose scaling forms for different quantities describing the network. In particular, when the distance effect is important (i) the connectivity distribution has a cut-off depending on the node density, (ii) the clustering coefficient is very high, and (iii) we observe a positive maximum in the degree correlation (assortativity) which numerical value is in agreement with empirical measurements. Finally, we show that if the number of nodes is fixed, the optimal network which minimizes both the total length and the diameter lies in between the scale-free and spatial networks. This phenomenon could play an important role in the formation of networks and could be an explanation for the high clustering and the positive assortativity which are non trivial features observed in many real-world examples.

cond-mat.dis-nn↗

arXiv2003

Connectivity Distribution of Spatial Networks

We study spatial networks constructed by randomly placing nodes on a manifold and joining two nodes with an edge whenever their distance is less than a certain cutoff. We derive the general expression for the connectivity distribution of such networks as a functional of the distribution of the nodes. We show that for regular spatial densities, the corresponding spatial network has a connectivity distribution decreasing faster than an exponential. In contrast, we also show that scale-free networks with a power law decreasing connectivity distribution are obtained when a certain information measure of the node distribution (integral of higher powers of the distribution) diverges. We illustrate our results on a simple example for which we present simulation results. Finally, we speculate on the role played by the limiting case P(k)=1/k which appears empirically to be relevant to spatial networks of biological origin such as the ones constructed from gene expression data.

cond-mat.dis-nn↗

arXiv2002

Large Scale Cross-Correlations in Internet Traffic

The Internet is a complex network of interconnected routers and the existence of collective behavior such as congestion suggests that the correlations between different connections play a crucial role. It is thus critical to measure and quantify these correlations. We use methods of random matrix theory (RMT) to analyze the cross-correlation matrix C of information flow changes of 650 connections between 26 routers of the French scientific network `Renater'. We find that C has the universal properties of the Gaussian orthogonal ensemble of random matrices: The distribution of eigenvalues--up to a rescaling which exhibits a typical correlation time of the order 10 minutes--and the spacing distribution follow the predictions of RMT. There are some deviations for large eigenvalues which contain network-specific information and which identify genuine correlations between connections. The study of the most correlated connections reveals the existence of `active centers' which are exchanging information with a large number of routers thereby inducing correlations between the corresponding connections. These strong correlations could be a reason for the observed self-similarity in the WWW traffic.

cond-mat.dis-nn↗

arXiv2002

Spatial Structure of the Internet Traffic

The Internet infrastructure is not virtual: its distribution is dictated by social, geographical, economical, or political constraints. However, the infrastructure's design does not determine entirely the information traffic and different sources of complexity such as the intrinsic heterogeneity of the network or human practices have to be taken into account. In order to manage the Internet expansion, plan new connections or optimize the existing ones, it is thus critical to understand correlations between emergent global statistical patterns of Internet activity and human factors. We analyze data from the French national `Renater' network which has about two millions users and which consists in about 30 interconnected routers located in different regions of France and we report the following results. The Internet flow is strongly localized: most of the traffic takes place on a `spanning' network connecting a small number of routers which can be classified either as `active centers' looking for information or `databases' providing information. We also show that the Internet activity of a region increases with the number of published papers by laboratories of that region, demonstrating the positive impact of the Web on scientific activity and illustrating quantitatively the adage `the more you read, the more you write'.

cond-mat.dis-nn↗

arXiv2002

The Random Fuse Network as a Dipolar Magnet

We introduce an approximate mapping between the random fuse network (RFN) and a random field dipolar Ising model (RFDIM). The state of the network damage is associated with a metastable spin configuration. A mean-field treatment, numerical solutions, and heuristic arguments support the broad validity of the approximation and yield a generic phase diagram. At low disorder, the growth of a single unstable `crack' leads to an abrupt global failure. Beyond a critical disorder, the conducting network sustains significant damage before the coalescence of cracks results in global failure.

cond-mat.stat-mech↗

arXiv2002

Truncation of power law behavior in "scale-free" network models due to information filtering

We formulate a general model for the growth of scale-free networks under filtering information conditions--that is, when the nodes can process information about only a subset of the existing nodes in the network. We find that the distribution of the number of incoming links to a node follows a universal scaling form, i.e., that it decays as a power law with an exponential truncation controlled not only by the system size but also by a feature not previously considered, the subset of the network ``accessible'' to the node. We test our model with empirical data for the World Wide Web and find agreement.

cond-mat.stat-mech↗

arXiv2000

Path Integral Approach to Strongly Nonlinear Composite

We study strongly nonlinear disordered media using a functional method. We solve exactly the problem of a nonlinear impurity in a linear host and we obtain a Bruggeman-like formula for the effective nonlinear susceptibility. This formula reduces to the usual Bruggeman effective medium approximation in the linear case and has the following features: (i) It reproduces the weak contrast expansion to the second order and (ii) the effective medium exponent near the percolation threshold are $s=1$, $t=1+κ$, where $κ$ is the nonlinearity exponent. Finally, we give analytical expressions for previously numerically calculated quantities.

cond-mat.dis-nn↗