Causal Effect Identification in lvLiNGAM from Higher-Order Cumulants

TL;DR

This paper advances causal inference in lvLiNGAM models by leveraging higher-order cumulants to identify causal effects even with limited or proxy instruments, supported by theoretical proofs and experimental validation.

Contribution

It introduces novel identifiability results and estimation methods for causal effects in lvLiNGAM with a single proxy or underspecified instruments, addressing key challenges in latent confounding.

Findings

Causal effects are identifiable with a single proxy or instrument.

Proposed methods outperform existing approaches in accuracy and robustness.

Theoretical proofs confirm identifiability under new conditions.

Abstract

This paper investigates causal effect identification in latent variable Linear Non-Gaussian Acyclic Models (lvLiNGAM) using higher-order cumulants, addressing two prominent setups that are challenging in the presence of latent confounding: (1) a single proxy variable that may causally influence the treatment and (2) underspecified instrumental variable cases where fewer instruments exist than treatments. We prove that causal effects are identifiable with a single proxy or instrument and provide corresponding estimation methods. Experimental results demonstrate the accuracy and robustness of our approaches compared to existing methods, advancing the theoretical and practical understanding of causal inference in linear systems with latent confounders.