Regensburg 2025 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

SOE: Fachverband Physik sozio-ökonomischer Systeme

SOE 10: Focus Session: Large Language Models, Social Dynamics, and Assessment of Complex Systems

Thursday, March 20, 2025, 15:00–18:15, H45

Recent Large Language Models (LLMs) learn the statistics of human communication, and hence they present a promising way of simulating societal interactions and opinion formation. Moreover, their flexible architecture makes them versatile to also learn the statistics of other complex systems. - We will discuss how to make best use of these advancements to model social interactions and dynamics across different scales and settings, how to use them to assess sentiments, narratives and opinions, and also how one can use LLMs to also understand other complex systems. Our aim is to offer an evaluative perspective on both the potential and challenges that are intrinsic to this rapidly evolving research area.

Organized by Vincent Brockers, Johannes Zierenberg, and Viola Priesemann

	15:00	SOE 10.1	Invited Talk: Emergent Behaviors in LLMs-Populated Societies — •Giordano De Marzo, Claudio Castellano, Luciano Pietronero, and David Garcia
	15:30	SOE 10.2	A Framework for Multi-Step Discussions of LLM-Based Agents — •Vincent Brockers, David Ehrlich, and Viola Priesemann
	15:45	SOE 10.3	Collective Turing Tests on LLMs — •Azza Bouleimen, Giordano de Marzo, Taehee Kim, Silvia Giordano, and David Garcia
	16:00	SOE 10.4	Dialogue React: Enhancing Conversation Synthesis for Social Science Simulations — •Ruggero Marino Lazzaroni
	16:15	SOE 10.5	Computational modeling of LLM powered personalized recommendations — •Alessandro Bellina, Giordano De Marzo, David Garcia, and Vittorio Loreto
	16:30	SOE 10.6	Statistical Modelling of Physics Classroom Interactions — •Niklas Stausberg, Karina Avila, Steffen Steinert, Jochen Kuhn, and Stefan Küchemann
	16:45		15 min. break
	17:00	SOE 10.7	Disentangling individual vs. collective contributions to polarization in political voting — Gavin Rees and •Edward Lee
	17:15	SOE 10.8	Understanding Information Spread in Social Networks through the Lense of Self-Organized Criticality: A Study on Telegram — •Roman Ventzke, Anastasia Golovin, Sebastian Bernd Mohr, Andreas Schneider, and Viola Priesemann
	17:30	SOE 10.9	Dynamics of information spread: the phase space of topics on Telegram — •Anastasia Golovin, Simon Bauer, Sebastian Mohr, and Viola Priesemann
	17:45	SOE 10.10	Computational approaches to analyzing political narratives — •Armin Pournaki and Tom Willaert
	18:00	SOE 10.11	Optimizing pandemic mitigation in the presence of seasonality, emerging variants and vaccination — •Laura Müller, Fabio Sartori, Jonas Dehning, Maximilian F. Eggl, and Viola Priesemann