Invariance, Causality, and Robustness

This lecture by the instructor from the University of Copenhagen focuses on the concepts of invariance, causality, and robustness in data analysis. The presentation covers topics such as instrumental variables, challenges in constructing tests for invariance, and the relationship between invariance and causality. The lecture also delves into the implications of invariance for distribution generalization and the role of interventions in modeling reality. Various challenges and questions related to invariance and causality are discussed, along with the proposal of choosing the best predictive model among invariant models. The session concludes with a summary emphasizing the link between invariance, causality, and robustness in data analysis.