Complex systems are characterized by high dimensionality, nonlinear interactions, stochastic influences, and emergent behaviour. They arise in a wide range of scientifically and societally relevant contexts, including biological and neuronal networks, adaptive and smart materials, epidemiological and climate systems, energy grids, and social dynamics.
Rapid advances in computing hardware, the growing availability of large and heterogeneous datasets, and recent developments in machine learning algorithms are turning artificial intelligence (AI) into a central tool of modern science. One the one hand, machine learning methods now enable new approaches across the entire scientific workflow, from experimental design and data acquisition to hypothesis generation, model discovery, and prediction. On the other hand, also emergent behaviour of AI is becoming the subject of investigations.
The Topical Programme “Artificial Intelligence and Complexity” brings together research on machine learning with the theory of complex systems, aiming at a systematic integration of data-driven methods into experiment, computation, and theory—especially in domains where classical approaches reach their limits. In particular, the programme explores the mutual reinforcement between AI and complexity science: insights from complex systems theory inspire new, interpretable, and robust machine learning methods, while AI-based approaches enable the analysis, modelling, and prediction of complex dynamic phenomena such as phase transitions, tipping points, and emergent collective behaviour. By combining data-driven techniques with domain knowledge—e.g. through physics-informed machine learning and model-based system identification—the programme fosters methodological innovation with broad applicability across disciplines. Furthermore, the programme investigates the optimization process itself as a complex dynamic system, characterizing the intricate topology of learning landscapes and the erratic or unstable learning trajectories that frequently emerge in high-dimensional settings.
The interdisciplinary Topical Programme combines the strong disciplinary expertise of various academic units of the University of Münster and is embedded within the Center for Data Science and Complexity (CDSC), providing an ideal framework for advancing fundamental research at the interface of artificial Intelligence and complex systems while enabling impactful applications in science and society.
Knowledge Base
A central structural element of the Topical Programme is the development of a digital interdisciplinary Knowledge Base. This resource is designed to lower entry barriers in interdisciplinary research by systematically curating key concepts, methods, data types, and modelling approaches from the participating disciplines. By providing a shared conceptual and methodological foundation—supported by interactive elements such as visualisations, glossaries, and executable examples—the Knowledge Base facilitates effective collaboration, accelerates scientific onboarding, and supports coherent training across disciplinary boundaries. It thus serves as a long-term enabling infrastructure for research and education at the interface of Artificial Intelligence and complexity science.
More information about the knowledgebase will be available soon.