Causal Inference with Deep Causal Graphs
This addresses the limitation of existing parametric causal models in handling real-life datasets with non-linear relationships, offering a method for more reliable causal estimations.
The paper tackles the problem of counterfactual estimation in causal inference by proposing Deep Causal Graphs and Normalizing Causal Flows, a neural network model that demonstrates expressive power in modeling complex interactions for applications like explainability and fairness.
Parametric causal modelling techniques rarely provide functionality for counterfactual estimation, often at the expense of modelling complexity. Since causal estimations depend on the family of functions used to model the data, simplistic models could entail imprecise characterizations of the generative mechanism, and, consequently, unreliable results. This limits their applicability to real-life datasets, with non-linear relationships and high interaction between variables. We propose Deep Causal Graphs, an abstract specification of the required functionality for a neural network to model causal distributions, and provide a model that satisfies this contract: Normalizing Causal Flows. We demonstrate its expressive power in modelling complex interactions and showcase applications of the method to machine learning explainability and fairness, using true causal counterfactuals.