Computational modelling of collective behaviour and cognition - Detailseite

This course will introduce students to computational models of collective behaviour and cognition. Such computational approaches can bridge the gap between the dynamics happening within and between agents in social groups, thereby mechanistically linking the individual and collective level. This ties together how (i) individual cognition drives social systems; (ii) the social environment drives individual cognition, and (iii) both levels interact. We will combine introductory lectures with practical hands-on sessions in R. We will start with a general introduction to computational models of cognition and typical Bayesian workflows including simulation, model fitting and parameter recovery. We then apply these methods to various computational models in social systems including models of (i) signal detection, (ii) evidence accumulation, and (iii) reinforcement learning. We will show how these approaches can help answer fundamental questions on social and collective decision making, such as when is it beneficial to learn from others (as opposed to on your own), and how do social learning strategies influence collective performance?

A basic understanding of programming is a requirement. The course will use coding exercises in the R environment, and a basic understanding of e.g. how to wrangle data in R, or another programming language, is required. By the end of the course, students are expected to have an overview over key concepts of Bayesian statistics and approaches to model decision making in social systems and will have developed step-by-step knowledge on the process of computational modelling of social systems.

Attendance is limited to 30 students. Please register for this course only if you seriously plan to take it.