Identifying Causal Effects Using a Single Proxy Variable

arXiv stat.ML / 4/13/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper addresses the difficulty of unobserved confounding in causal inference by assuming access to a single proxy variable for the latent confounder.
It introduces the SPICE framework, which proves identifiability of causal effects under a completeness assumption about the known mechanism generating the proxy from the confounder.
The authors generalize prior proxy-based identifiability results to allow multi-dimensional proxies, more flexible functional forms, and a wider range of distributional settings.
They propose SPICE-Net, a neural-network-based estimation method that supports both discrete and continuous treatments.

Abstract

Unobserved confounding is a key challenge when estimating causal effects from a treatment on an outcome in scientific applications. In this work, we assume that we observe a single, potentially multi-dimensional proxy variable of the unobserved confounder and that we know the mechanism that generates the proxy from the confounder. Under a completeness assumption on this mechanism, which we call Single Proxy Identifiability of Causal Effects or simply SPICE, we prove that causal effects are identifiable. We extend the proxy-based causal identifiability results by Kuroki and Pearl (2014); Pearl (2010) to higher dimensions, more flexible functional relationships and a broader class of distributions. Further, we develop a neural network based estimation framework, SPICE-Net, to estimate causal effects, which is applicable to both discrete and continuous treatments.