Causal Representation Learning: Optimizing Representations to Preserve Interventional Control

Causal representations that preserve interventional control over a target variable while minimizing information about the input variables.