Dr. Malte Vogl (he/him)

Knowledge spread in complex socio-technical systems

Core theme: The technosphere as a complex system,  with J. Büttner, G. Steudle

Taking the evolution of scientific disciplines as inspiration, we build an Agent-based model that takes into account the growth of the numbers of actors, together with the ageing of agents. Agents have incentives to explore and create the knowledge landscape with the aim to get academic credit. In this model we want to observe the scaling effects a growing number of actors has on the dynamics of knowledge spread. Simulation experiments are run, e.g. by initializing the agents with different initial epistemic positions (opposing, centralized, fragmented) and observe the emergence or not of long term steady states.

In a second step, we introduce conferences as extreme events in the knowledge spreading dynamics. In the formative phase of new disciplines,  conferences can bring together agents from diverse epistemic positions, whose dynamics usually not interact, and last only a limited amount of simulation time.

• Emergent potential landscapes and random searches
• Gatherings at the begining of the formation of disciplines as extreme events
• Critical transitions in connectivity
• Global knowledge connection and the future of innovation

Challenges of modeling societal growth and disruption

Core theme: Impacts of extreme events and other shocks, with G. Steudle, J. Donges

The above introduced agent-based model hints at some more fundamental challenges in the modelling of artificial societies. We develop best practices to approach these, by setting up well-described and documented minimal models that incorporate birth and death-like processes of agents. Furthermore, we approach the fundamental problem of the modeling of disruption in these agent-based models. While it is simple to introduce exogenous parameter changes that disrupt dynamics, setting up agents behaviours such that disruption can emerge endogenously is much more challenging. We aim at finding a set of simple enough interaction rules, such that a mathematical reduction of the system dynamics remains possible.

• Minimal model to incorporate growing number of agents with scaling of connectivity
• Basics of shock modelling: exogenous vs endogenous approaches

Model systems for knowledge and earth-system dynamics feedback

Core theme: Anthropocene engine, with G. Steudle, J. Büttner

Another line of explorative model generation focuses on the time-scales of feedback between knowledge and earth system dynamics. The influence of the technosphere on earth-system dynamics is evident. The accelerating feedback between knowledge and technology, e.g. in the generation of new AI methods, coincides with more frequent extreme weather events. Seeing this as the convergence of different timescales allows to look at the human-technology-earth system from the angle of time-delayed feedback control and complexity science. We aim to understand the aspect of dual instabilities, both in the earth and the human societal system.

• Aspects of time-delayed feedback control
• Time-scale convergence vs separation and its effects on stability

Learning and outreach for complex system thinking

Core theme: Great Acceleration Observatory, with R. Winkelmann, P. Roberts, A. Kaye

Acting in the increasingly unstable human-earth system requires a fundamental knowledge of complex system thinking, including the capability to self-reflect. In this sub-project we aim to create communicative devices that allow to transport complex information while staying accessible for a wide audience. This includes interactive visualizations, immersive 3D interfaces and serious games. Each development is accompanied with sufficient documentation and tutorials to make it a reusable resource for all research at the institute.

• Visualizations of temporal change
• Games as communicative devices

Fundamentals of data integration and data biases

Core theme: Great Acceleration Observatory, with A. Kaye

Communicating scientific results of the institute can lead to strong reactions of audiences, both positive and negative. Together with common scientific practices of e.g. research data preservation, this is a strong motivation to asses the role data and model generation and availability has on what we are able to think, both about events of the past as well as projections of possible futures. We therefore develop methods for the simulation of missing or erroneous data on modelling outcomes and establish institute-wide best-practices for code review and software or data publications.

• Epistemics of simulations
• Archival biases, accessing new sources
• Open Science, open software
• Peer-review and pre-publishing for data and code