The purpose of the research project is to answer how and to what extent causal effects can be estimated from incomplete data. The focus is on partially observed dynamical systems.
The background is the data revolution in science, industry and business, where empirical and data driven models are taking over decision making. Examples, where model building from first principles is difficult, include the biological cell, the brain, traffic systems and social networks. Causal questions are the most important to answer: what happens if we intervene and how can we intervene to achieve a goal?
The mathematical language of causality is well developed and based on probabilistic graphical models, but the methodology for inferring causal models is incomplete. This is particularly so for dynamical systems or when some aspects of the system are unobservable.
The project is funded by VILLUM FONDEN.
|Niels Richard Hansen||Professor, P.I.|
|Steffen Lauritzen||Professor, Co-P.I.|
|Jonas Peters||Associate Professor, Co-P.I.|
|Søren Wengel Mogensen||Ph.D.-student|