This paper introduces a novel approach to counterfactual inference using quantile regression, establishing a connection between counterfactual outcomes and quantiles. The proposed method offers efficient and effective estimation of counterfactual outcomes without the need for a predefined structural causal model.
This paper introduces a novel approach to counterfactual inference using quantile regression, providing reliable predictions without the need for structural causal models. By reframing counterfactual inference as an extended quantile regression problem, the method offers superior statistical efficiency and generalization capabilities.