Network Science for Multidimensional Data Analysis
The coexistence of multiple types of interactions within social, technological and biological networks has moved the focus of the statistical physics of complex systems towards their description as a set of subsystems organized as layers of connectivity. This novel approach has unveiled that the multilayer nature of complex systems has strong influence in the emergence of collective states and their critical properties.
Although recently spurred by the burst of datasets in which different means of interactions within the same system are encoded, the interest in the multilayered nature of complex systems dates back several decades and span across diverse disciplines. This has enhanced the activity devoted to multilayered networks of many network science practitioners during the recent years and, nowadays, the topic is one the most important directions in the field.
A necessary step to improve our understanding of multilayer complex networks is to generalize models and tools of traditional network theory and to develop a unified framework covering all the possible types of interdependence between network layers. A natural description of multilayered systems consists in a population of nodes having different sets of neighbors in each layer (which can represent, e.g., a task, an activity, or a category). A fundamental aspect controlling the collective behavior of multilayer networks is how the interaction between the different types of connections is set. In particular, this type of (interlayer) interactions are intrinsic of multilayer networks, and it has been shown to be responsible for the emergence of novel collective phenomena due to the new structural and dynamical correlations between the components of a system that they introduce.
The scope of the satellite meeting is to review the recent advances in the field of multilayer (both edge-colored multi-graphs and interconnected) networks, focusing in particular on the interplay between structure and dynamics. We will pay special attention to novel applications in which the multilayer formulation is crucial. The list of topics that we aim to cover at the conference is the following:
1. Mathematical properties of multilayer and interconnected structures
2. Empirical measurements of multilayer and interconnected networks
3. Applications of such models to biological, social, economic, technological and urban systems
4. The general framework of Data Science of Multilayer Networks
Universidad de Zaragoza
|8 January 2018
|Call for abstracts
|15 March 2018
|Deadline for abstract submission
|1 April 2018
|Notification of acceptance
|11 June 2018
Opening and Chair
Multilayer connectomes: new methodologies for studying brain networks
Cities' centrality according to the changing space of their firms' activities (2010-2016)
Multiplex model of mental lexicon reveals explosive learning in humans
We propose a multiplex network representation of the mental lexicon of word similarities as a natural framework for investigating large-scale cognitive patterns. Our findings provide quantitative confirmation of existing conjectures about the emergence and presence of core structure in the mental lexicon during cognitive development.
Determinants of public cooperation in multiplex networks
Recent works stated that multilayer networks enhance cooperation through interdependent network reciprocity. While it is true that multiplexity can sustain cooperation under adverse conditions, our research shows that this is possible only when some structural and dynamical constrains are fulfilled. These results caution against early overly optimistic predictions that the presence of multiple social domains may in itself promote cooperation.
Diverse types of percolation transitions induced by cluster merging dynamics in restricted percolation models with two groups
We introduce a restricted percolation model, in which clusters are classified into two sets composed of a portion of clusters of smallest sizes and the rest at each dynamic step. The clusters in the former set are encouraged to merge to another cluster, while the rest clusters are discouraged. This simple rule creates a hybrid percolation transition with continuous varying critical behavior in static networks and an abnormal type of percolation transition in growing networks.
Principled parameter selection for modularity maximization in multilayer networks
One way to detect communities in multilayer networks is by maximizing a generalized version of the well-known modularity function. A shortcoming of modularity is that it involves two sets of tuning parameters that have been difficult to choose in practice. We show how to select these parameters in a principled way for various kinds of multilayer networks including temporal, multiplex, and multilevel (i.e., hierarchical) networks.
Metric projection for dynamic multiplex networks
A novel approach is introduced here for the longitudinal analysis of a time series of multiplex networks, defined by mean of a metric transformation conveying the information carried by all layers into a single network for each timestamp, with the original layers as nodes. The transformation is induced by the Hamming–Ipsen–Mikhailov distance, preserving the key events encoded into each instance of the multiplex network time series.
Epidemic Spreading on Multiplex Networks
Multirelational and multilayer networks have been used in the social sciences for more than thirty years to represent different types of relations between social entities. However, only in the last few years, and fuelled by huge amounts of new data, "The Multiplex" conquered other fields, finally reaching the complex systems and network science communities. The impact of this new modelling approach to all areas of network science has been tremendous with the discovery of many new phenomena, not seen in single-layered systems. This is particularly true for the study of spreading processes where hundreds of works have been produced in less than ten years. With this talk I will try to put some order in this new "Wild West" and show how the multiplex and multilayer paradigms helped solve several longstanding questions in mathematical epidemiology. I will start by presenting how to extend classical spreading dynamics to the multiplex case, and the new phenomena that arise from there. Then, I will continue by showcasing how the multiplex framework has allowed to tackle unresolved problems like: the interaction between different diseases, epidemic modelling on temporal networks and, finally, including distinct mobility patterns in large-scale models.
Peter J Mucha
CHAMP: Post-Processing Partitions for Modularity Optimization
We present the Convex Hull of Admissible Modularity Partitions (CHAMP) algorithm, pruning and prioritizing partitions identified across multiple community detection results at different resolution parameters and, for multilayer networks, different interlayer coupling weights.
Multiplex Network Regression: a Statistical Framework for Multiplex Network Analysis
In this talk, we propose a new nonlinear parametric model to perform statistical regression on networks. The method is based on an extension to multiplex networks of the generalised hypergeometric ensembles. Thanks to the features provided by the ensembles, the method allows to quantify the influence that each layer in a multiplex network has on the observed interactions between nodes.
Multilayer network based mining of Linked Open Data and its application to multi-criteria university ranking
We transform subject–predicate–object triple-sets of RDF data models into multilayer networks. We define the edge weights according to the subjective importance of the predicates of the RDF data models and calculate the node centralities for multi-objective ranking. The DBpedia based analysis of the cultural impact of higher education institutes demonstrates the applicability of the proposed methodology.
A network approach to airports mobility
We analyze the mobility behaviors of workers and travellers inside airports facilities by applying temporal and multilayer networks metrics on a large dataset of anonymised GPS trajectories. This approach shows the opportunity of using these tools for the analysis of spatial trajectories, deepening our knowledge on how humans interact in public spaces.
Cascade propagation in random and spatial interdependent networks
When the functionality of a node in a complex network requires the simultaneous functioning of a node in another network, even a single failed node can lead to a cascade of failures throughout the system[1,2]. This cascade has distinctive properties which are determined by the network topology, including nucleation transitions and a metastable phase in spatial networks[3,4]. In this talk we will review recent advances on interdependent spatial network cascades including spatially localized attacks [5,6] and the impact of tunable embedding strength . We will compare these results with cascades in random networks and the spatio-temporal patterns arising from dynamical interdependence . These distinctive cascade patterns shed light on the impact of interdependence in multilayer networks.
Manlio De Domenico