keynote speakers

 
lada_adamic.png

Lada Adamic
Facebook Inc.

reka_albert.png

Reka Albert
Pennsylvania State University, USA

urlik_brandes.png

Ulrik Brandes
ETH Zürich, Switzerland

 
jari_saramaki.png

Jari Saramäki
Aalto University, Finland

stefan_thurner.png

Stefan Thurner
Medical University of Vienna, Austria

michael_varizgiannis.png

Michalis Vazirgiannis
LIX, École Polytechnique, FRance

reka_albert.png

 Reka Albert

Penn State University

Prof. Réka Albert received her Ph.D. in Physics from the University of Notre Dame (2001), working with Prof. Albert-László Barabási, then did postdoctoral research in mathematical biology at the University of Minnesota, working with Prof. Hans G. Othmer. She joined Penn State in 2003, where she currently is a Distinguished Professor of Physics with adjunct appointments in the Department of Biology and the Huck Institute of the Life Sciences. Prof. Albert is a network scientist who works on predictive modeling of biological regulatory networks at multiple levels of organization. Dr. Albert's pioneering publications on the structural heterogeneities of complex networks had a large impact on the field, reflected in their identification as "Fast breaking paper" and "High impact paper". Prof. Albert is a fellow of the American Physical Society and of the Network Science Society and an external member of the Hungarian Academy of Sciences. She was a recipient of an NSF Career Award (2007), the Maria Goeppert-Mayer award (2011), and the Distinguished Graduate Alumna Award of the University of Notre Dame (2016). Her service to the profession includes serving on the editorial board of the Biophysical Journal, Bulletin of Mathematical Biology, npj Systems Biology and Applications, and as peer reviewer for more than 35 journals.

Network-based dynamic modeling of biological systems: toward understanding and control

My group is using network science to understand the emergent properties of biological systems. As an example, we think of cell types as attractors of a dynamic system of interacting (macro)molecules, and we aim to find the network patterns that determine these attractors. We collaborate with wet-bench biologists to develop and validate predictive dynamic models of specific systems. We then use the specific knowledge gained to draw general conclusions that connect a network's structure and dynamics. An example of such a general connection is our identification of stable motifs, self-sustaining cyclic structures in the network that determine a trap subspace of the system’s state space, or equivalently determine points of no return in the dynamics of the system. We have shown that control of stable motifs can guide the system into a desired attractor. Such attractor control can form the foundation of therapeutic strategies on a wide application domain. I will illustrate such applications in our model of a cell fate change that represents the first step toward cancer metastasis. Several model-predicted therapeutic interventions to block this cell fate change were validated experimentally.

urlik_brandes.png

Urlik Brandes

ETH Zürich

Ulrik Brandes is a professor of social networks at ETH Zurich since 2017. With a background in algorithmics, his main interests are in network analysis and visualization, with application to social networks in particular. He is a co-author of the visone software for network analysis and of the GraphML data format. Deutsche Forschungsgemeinschaft (DFG) awarded him a Reinhart Koselleck-Project on Social Network Algorithmics, in which he took a shot at improving the methodological foundations of network science, and he was a principal investigator in the ERC Synergy Project NEXUS 1492 where he worked on reconstructing archaeological networks from fragmented and heterogeneous observations.
Brandes received a Diploma degree from RWTH Aachen in 1994and a PhD from the University of Konstanz in 1999, both in computer science.
After postdoctoral research visits to Brown University and the University of Sydney, he completed his habilitation in 2002 and became associate professor at the University of Passau the same year. From 2003-2017 he was full professor of algorithmics at the University of Konstanz.
He is a member of the board of directors of the International Network for Social Network Analysis (INSNA) since 2008, and was a member of the Graph Drawing Steering Committee 2007-2014. He acts as the coordinating editor of Network Science and as an associate editor of Social Networks, and he is an editorial board member of the Journal of Mathematical Sociology as well as the Journal of Graph Algorithms and Applications.

On a Positional Approach to Network Science

This presentation is about network science methodology. By viewing it as a data science rather than, say, a collection of methods or a unifying theory, we create opportunities for more rigorous research, both mathematically and empirically. Pivotal to the adaptation of methods to general, multivariate and temporal, situations is the notion of network position, which summarizes the relationships of a node with the rest of the network. I will give examples showcasing how the analysis of centralities, roles, and communities can benefit from a positional perspective.

jari_saramaki.png

Jari Saramäki

Aalto University, Finland

Jari Saramäki is a full professor and vice head at the Department of Computer Science, Aalto University, Finland. He received his PhD in applied physics in 1998, studying quantum crystals at milliKelvin temperatures. After some career twists and turns involving technology companies and what we would nowadays call data science, he returned to academia in 2003 to study complex networks, a new and rapidly expanding field at that time. Jari Saramäki is probably best known for his work on social and temporal networks, but his broad range of research interests has included topics from ant supercolonies to the human immune system. 

Temporal networks: past, present, future

The key strength of network science has been its ability to strip away unnecessary details, making it easier to grasp the inner workings of systems that are large and complex. At the same time, however, entire subfields have emerged that build on adding back some of this detail: weighted networks, multilayer networks, and temporal networks, the latter being the topic of this talk. I will provide an overview of what temporal networks are and what the temporal networks framework can do, and discuss when the temporal-network treatment is useful and when not. I will discuss some key findings and methods, using time-stamped social interactions as an example case, and finally, try to sketch some future directions for temporal-network research.

stefan_thurner.png

Stefan Thurner

Medical University of Vienna, Austria

Stefan is full professor for Science of Complex Systems at the Medical University of Vienna. He is the president of the Complexity Science Hub Vienna, external professor at the Santa Fe Institute, and a senior researcher at IIASA. Stefan obtained a PhD in theoretical physics from the Technical University of Vienna and a PhD in economics from the University of Vienna. Stefan started his career in theoretical particle physics and gradually shifted his focus to the understanding of complex adaptive systems. He published about 200 articles in physics, applied mathematics, network theory, evolutionary dynamics, life sciences, economics and finance, and lately in social sciences. He holds two patents. His work has been covered by international media such as the New York Times, BBC world, Nature, New Scientist, Physics World, and is featured in more than 400 newspaper, radio and television reports. Stefan was elected Austrian “scientist of the year” in 2018.

How to eliminate systemic risk from financial multi-layer networks

Given the detailed network structure of financial obligations in financial markets one can compute not only compute the systemic risk contribution of the individual financial players, but also it becomes possible to estimate the contribution of systemic risk of every single financial transaction. This in turn allows us to design incentive schemes for market participants to become systemic risk sensitive, by preferring systemically unrisky transactions. We show that such schemes lead to a restructuring of financial exposure networks in ways that suppress the possibility of cascading failure and thereby drastically reduces systemic risk. We discuss ways to compute optimal financial networks that can be used to benchmark and monitor actual financial networks.

michael_varizgiannis.png

 Michalis Vazirgiannis

Ecole Polytechnique, France

Dr. Vazirgiannis is a Professor at LIX, Ecole Polytechnique in France. He has conducted research in Frauenhofer and Max Planck-MPI (Germany), in INRIA/FUTURS (Paris). He has been a teaching in AUEB (Greece), Ecole Polytechnique, Telecom-Paristech, ENS (France), Tsinghua, Jiaotong Shanghai (China) and in Deusto University (Spain). His current research interests are on deep and machine learning for Graph analysis (including community detection, graph classification, clustering and embeddings, influence maximization), Text mining including Graph of Words, deep learning for word embeddings with applications to web advertising and marketing, event detection and summarization. He has active cooperation with industrial partners in the area of data analytics and machine learning for large scale data repositories in different application domains. He has supervised twenty completed PhD theses. He has published three books and more than a 200 papers in international refereed journals and conferences and received best paper awards in ACM CIKM2013 and IJCAI2018. He has organized large scale conferences in the area of Data Mining and Machine Learning (such as ECML/PKDD) while he participates in the senior PC of AI and ML conferences – such as AAAI and IJCAI. He has received the ERCIM and the Marie Curie EU fellowships, the Rhino-Bird International Academic Expert Award by Tencent and between 2015 and 2018 he lead the AXA Data Science chair.

Machine learning for Graphs based on Kernels

Graph kernels have attracted a lot of attention during the last decade, and have evolved into a rapidly developing branch of learning on structured data. During the past 20 years, the considerable research activity that occurred in the field resulted in the development of dozens of graph kernels, each focusing on specific structural properties of graphs.  Graph kernels have proven successful in a wide range of domains, ranging from social networks to bioinformatics. The goal of this presentation is to provide a unifying view of the literature on graph kernels.  In particular, we present a comprehensive overview of a wide range of graph kernels.  Furthermore, we perform an experimental evaluation of several of those kernels on publicly available datasets, and provide a comparative study. Finally, we discuss key applications of graph kernels, and outline some challenges that remain to be addressed. The experimental comparison was based on an open source python library (Grakel) we designed implementing all the known so far graph kernels.