Counterfactual Identifiability via Dynamic Optimal Transport

ArXi:2510.08294v2 Announce Type: replace-cross We address the open question of counterfactual identification for high-dimensional multivariate outcomes from observational data. Pearl argues that counterfactuals must be identifiable (i.e., recoverable from the observed data distribution) to justify causal claims. A recent line of work on counterfactual inference shows promising results but lacks identification, undermining the causal validity of its estimates.