Invited Speakers

Renaud Lambiotte

Bio: Renaud Lambiotte is professor of Networks and Nonlinear Systems at the University of Oxford. He received his PhD in theoretical physics from Université libre de Bruxelles in 2004, and has been a Research Associate at ENS Lyon, Université de Liège, Université catholique de Louvain and Imperial College London, and professor of Mathematics at the University of Namur. His research interests include network science, data mining, stochastic processes, social dynamics and neuroimaging. He has published around 130 peer-reviewed articles and one book on temporal networks. He is also the co-founder of L’Arbre de Diane, a publishing company at the interface between science and literature.

Website: https://www.maths.ox.ac.uk/people/renaud.lambiotte

Talk title: Structural balance beyond signed networks

Abstract: Structural balance—whether in its weak or strong form—is a foundational concept in the theory of signed networks. In this talk, I will explore how this notion extends beyond its traditional setting to offer insight into more general classes of networks. In particular, I will demonstrate that structural balance provides a natural framework for analyzing complex-weighted networks, where edge weights are complex numbers, introducing phenomena such as signal rotation and interference. I will show how structural balance can be used to classify these networks and to describe the behavior of random walks on them. Notably, asymptotic local consensus emerges when the network is structurally balanced, while global consensus arises in strictly unbalanced cases. Finally, I will present a broader generalization in which signals propagate through nodes via arbitrary linear transformations associated with each edge, moving the discussion into higher-dimensional dynamics.

Jana Diesner

Bio: Jana Diesner is a Full Professor at the Technical University of Munich, School of Social Science and Technology, Department of Governance, with an affiliate appointment at the School of Computing, Information and Technology. She leads the Human-Centered Computing group. Her interdisciplinary group works on methods from network analysis, natural language processing, machine learning and AI, and integrates them with theories from the social sciences and humanities to advance knowledge about socio-technical systems and responsible computing. Jana earned her Ph.D. at Carnegie Mellon, School of Computer Science. Jana joined TUM from the University of Illinois Urbana Champaign, where she was a tenured professor at the School of Information Sciences. There, she served as the Director of the Ph.D. Program, Director of Undergraduate Programs, and Director of Strategic Initiatives/Data Science.

Website: https://www.gov.sot.tum.de/hcc/team/jana-diesner/

Talk title: Natural language processing to extract and enhance network data

Abstract: Data on edge valence is hard to collect. For example, the reciprocity of friendship ties when elicited with questionnaires or surveys, is a little over 50%. This asks for alternative methods for data collection methods. I show how we have been using methods from natural language processing to identify edge signs, and how we used this approach to re-evaluate balance theory. I also show how biases in natural language processing methods lead to biased network analysis results.

Samin Aref

Bio: Samin Aref is an Assistant Professor, Teaching Stream in Data Science at the University of Toronto. Prof. Aref holds a PhD in Computer Science from the University of Auckland (New Zealand, 2019) and an MSc in Industrial Engineering from Sharif University of Technology (Iran, 2014). Prior to joining the faculty at the University of Toronto, he has been a research scientist and a research area chair at the Max Planck Institute for Demographic Research, Laboratory of Digital and Computational Demography (Germany, 2018-2021). Prof. Aref’s areas of interest are Network Science, Machine Learning, Optimization, Artificial Intelligence, Economics, and Social Computing for which he has secured funding from several organizations in Germany, New Zealand, and Canada. Prof. Aref has developed four new courses on these topics which he teaches to over 700 students annually at the University of Toronto. He has delivered more than 90 invited and contributed talks at international conferences and workshops and has authored over 30 papers in reputable journals and conference proceedings.

Website: https://saref.github.io/

Talk title: Integer Programming Models for Optimization-Based Clustering and Analysis of Networks

Abstract: A signed network is one with positive and negative edges. The structure of signed networks can be analyzed from the perspective of balance theory. A signed network is strongly balanced (weakly balanced) if its nodes can be partitioned into k≤2 clusters (k clusters) such that positive edges are within the clusters and negative edges are between the clusters. We propose new integer programming models to optimally partition network nodes by minimizing the total number of intra-cluster negative and inter-cluster positive edges. These optimization models partition a network into clusters of nodes that are the most fitting to the balance theory and therefore reveal the extent of alignment between balance theory and any input network (partial balance). We measure the distance of a network to weak and strong balance using the minimum number of edges whose removal leads to weak and strong balance respectively. The concepts of strong and weak balance in signed networks can be extended to applications beyond the classic friend-enemy interpretation of balance theory in the social context. Through extensive computational analysis, we explore common structural patterns across a wide range of networks. Exact optimization is also relevant for the clustering of unsigned networks; a case that we make by comparing 30 community detection algorithms. This talk provides an overview of using integer programming to develop exact optimization-based clustering algorithms for signed and unsigned networks. In collaboration with: Mark C. Wilson (University of Massachusetts Amherst), Andrew Mason (University of Auckland), Zachary P. Neal (Michigan State University), Ly Dinh (University of South Florida), Rezvaneh Rezapour (Drexel University), Jana Diesner (Technical University of Munich), Mahdi Mostajabdave (Polytechnique Montréal), Hriday Chheda (University of Toronto) and Boris Ng (University of Toronto)