Counterfactual Identifiability of Bijective Causal Models

TL;DR

This paper investigates counterfactual identifiability in bijective causal models, providing theoretical results and practical methods for learning such models, with applications in visual tasks and video streaming simulations.

Contribution

It introduces a new class of causal models called bijective generation mechanisms and develops a practical learning approach for counterfactual inference in these models.

Findings

Established counterfactual identifiability for models with unobserved confounding

Proposed a structured generative modeling approach for learning BGMs

Demonstrated effectiveness in visual and video streaming tasks

Abstract

We study counterfactual identifiability in causal models with bijective generation mechanisms (BGM), a class that generalizes several widely-used causal models in the literature. We establish their counterfactual identifiability for three common causal structures with unobserved confounding, and propose a practical learning method that casts learning a BGM as structured generative modeling. Learned BGMs enable efficient counterfactual estimation and can be obtained using a variety of deep conditional generative models. We evaluate our techniques in a visual task and demonstrate its application in a real-world video streaming simulation task.