Facundo Fainstein

Pls RT this ad: I have an opening for a PhD student in my @IntBioPhysics group to begin ASAP , part of a collaboration with the group of Ilya Nemenman @INemenman at Emory university. Deadline: 1st November Details here: stellen.uni-konstanz.de/jobposting/099… and application portal here: stellen.uni-konstanz.de/en/jobposting/… We are seeking a doctoral student with a quantitative background. Ideally, the candidate should have a master degree in physics but other quantitative backgrounds will be considered. The researcher will be based in the Integrative Biophysics group at the University of Konstanz and Max Planck Institute of Animal Behaviour, located in Konstanz, Germany. The project summary is as follows: Project Title: Inferring dynamical interactions in social groups Project Abstract: Social interactions are one of the most foundational building blocks of group behaviors, shaping decision making strategies across ecologies. Social groups consist of small number of individuals in contrast with collectives that consist of large number of agents. Despite the importance of social groups as a building block of living systems, there is little understanding of mechanistic details of the nature of those social interactions and how they shape emergent social functional properties. This stands in contrast with physics of collectives, which usually focus on the dynamics of very large group of individuals and have been thoroughly studied. In an experimental system or in field observations tackling social behaviors, we do not know a priori the social forces, nature of those interactions and their dynamics. Recent advances in software and hardware allow us to track the behavior of multiple individuals over extended spatio-temporal scales. Thus it is crucial to develop methodologies to infer dynamical social interactions and social forces from experimental observations. This relates closely to work done inferring "social forces" from dusty plasmas. Recent theoretical work has developed an information theoretic approach to estimate intrinsic motivation, based on maximizing an agent's empowerment (defined as the mutual information between its past actions and future states). This approach can be used to study social behaviours, where individuals choose actions without an explicit reward signal. In this project, we are planning to apply this framework to infer and understand social behaviour of small animal groups specifically locusts, mice and marmosets monkeys offering not only a mechanistic insights into mechanisms of social organization but also make quantitative evolutionary comparison of social behaviours. 1. Yu, Wentao, Eslam Abdelaleem, Ilya Nemenman, and Justin C. Burton. "Physics-tailored machine learning re-veals unexpected physics in dusty plasmas.Proceedings of the National Academy of Sciences 122, no. 31 (2025): e2505725122. 2. Tiomkin, Stas, Ilya Nemenman, Daniel Polani, and Naftali Tishby. "Intrinsic motivation in dynamical control sys-tems.” PRX Life 2, no. 3 (2024): 033009.

A machine learning method allows reconstructing attractors from data, ensuring that the topology of the underlying phase space is preserved. @Exactas_UBA @Fisica_UBA @fainsteinfacu @GaboMindlin @pgroisma aippub.org/4hflQeu