We are interested in complex systems and mathematical tools to model and study them.
Complex systems are ubiquitous in our lives and the amount of available data that describe complex systems has increased tremendously over the past years. Some well-known examples of complex systems include social networks, public transport networks, and networks of financial transactions.
In the complex systems cluster, we are interested in different aspects of complex systems. First, modelling: what is the right way to represent complex systems and how can we model them faithfully? Second, analysis: what methods can we use to analyse the models of complex systems and what do they tell us about how the systems work? And third, interpretation: how can we interpret our findings and put them into context?
Our research projects are diverse, ranging from methods for community detection through modelling and knowledge representation to applications in finance, technology, and biology. But what brings us all together here is our interest in graph theory and network science, and the opportunity to exchange ideas.
Check out our activities at the bottom of the page. If you want to get in touch, you can contact the cluster leader directly, or all cluster members by sending an email to our mailing list: ctc_complexsystems [at] wasp-sweden [dot] se.
Cluster Leader
Cluster Members
Vilhelm Agdur
Uppsala University, Department of Mathematics,
supervised by Fiona Skerman
Keywords: community detection, random graphs
Interests in discrete probability, particularly random graphs and combinatorial optimisation problems on random graphs, and algorithmic properties of such problems. Methods from statistical physics have shown themselves to be useful in this area, so my interests tend to fall in the intersection between discrete probability, computer science, and statistical mechanics.
Email: vilhelm.agdur [at] math.uu [dot] se
Christopher Blöcker
Umeå University, Department of Physics
supervised by Martin Rosvall
Keywords: community detection, random walk
I am interested in community detection and study how the information-theoretic framework known as the map equation can be extended to incorporate more domain-specific information to improve the quality of results.
Email: christopher.blocker [at] umu [dot] se
Chalmers, Department of Biology and Biological Engineering
Keywords: systems biology, knowledge graphs, automation, active learning
KTH, Division of Robotics, Perception, and Learning
supervised by Florian Pokorny
Keywords: graph generation, weighted and attributed graphs, machine learning, comparing graphs, spectral graph theory
My research relates to advanced analytics for anti money laundering. Specifically, I’m interested in how recent advanced in machine learning and topological data analysis can be applied to financial transaction data for discovery of money laundering schemes. Each year a vast amount of dirty money is laundered through the financial systems and financial instiutions are under pressure to address this. The problem of discovering such schemes nevertheless remains very challenging from a machine learning perspective due to lack of labelled data, large data volumes and data secrecy.
KTH, Division of Software and Computer Systems
My research interests primary lie in leveraging topological data analysis (TDA) to improve interpretation, transparency and results of language models (NLP).
Uppsala University, Department of Mathematics
Chalmers, Department of Biology and Biological Engineering
Lund University, Department of Automatic Control
Keywords: robustness and sensitivity analysis, random graphs, pseudospectrum, non-normal operators
KTH, Department of Mathematics
My interests are in enumerative, algebraic and geometric combinatorics and their role in the mathematical foundations of artificial intelligence.
KTH, Division of Theoretical Computer Science
supervised by Aristides Gionis
Keywords: mobility Prediction, point-of-interest, social network
Lund University, Department of Mathematics
Keywords: graph rigidity, graph embedding, distance geometry, graph signal processing
KTH, Division of Robotics, Perception, and Learning
supervised by Florian Pokorny
My research interests lie in the area of artificial intelligence, with a particular focus on applications of geometric and topological methods to analyze or improve the methods of machine learning.
Lund University, Department of Computer Science
supervised by Görel Hedin
Linköping University, Department of Computer and Information Science
Cluster Activities
We met for the first time to organise and discuss how we run the cluster.
We met and discussed our research projects and interests.
Daniel presented his research project and the cluster provided feedback.
Ciwan presented his research project and the cluster provided feedback.
We discussed the paper The why, how, and when of representations for complex systems to establish a common vocabulary.
Alexander presented his research project and the cluster provided feedback.
We discussed the paper When is a Network a Network? Multi-Order Graphical Model Selection in Pathways and Temporal Networks and how it relates to our research.
Together, we are taking the Stanford course Machine Learning with Graphs, taught by Jure Leskovec and available on YouTube, and discuss the content. This time, lectures 1-6.
Together, we are taking the Stanford course Machine Learning with Graphs, taught by Jure Leskovec and available on YouTube, and discuss the content. This time, lectures 7-8.
Together, we are taking the Stanford course Machine Learning with Graphs, taught by Jure Leskovec and available on YouTube, and discuss the content. This time, lectures 9-10.
Filip presented a research question and we discussed.
We reconnected after summer and Christopher presented his recent results from his project Mapping flows on weighted and directed networks with incomplete observations.
Jonas presented his research project and the cluster provided feedback.
Martin presented his research project and the cluster provided feedback.
In connection with the Winter Conference, we held a virtual poster session on Mozilla Hubs.
We met and discussed our recent research progress.