Towards Robust Causal Effect Identification Beyond Markov Equivalence

TL;DR

This paper introduces a new criterion for identifying causal effects using background knowledge, even when the causal graph cannot be precisely determined due to untestable assumptions or limited data.

Contribution

It provides a sufficient condition for causal effect identification from multiple Markov equivalence classes with background knowledge, extending beyond traditional assumptions.

Findings

New criterion for causal effect identification

Applicable when causal graph is not uniquely identifiable

Enhances robustness in causal inference

Abstract

Causal effect identification typically requires a fully specified causal graph, which can be difficult to obtain in practice. We provide a sufficient criterion for identifying causal effects from a candidate set of Markov equivalence classes with added background knowledge, which represents cases where determining the causal graph up to a single Markov equivalence class is challenging. Such cases can happen, for example, when the untestable assumptions (e.g. faithfulness) that underlie causal discovery algorithms do not hold.