Identifying Causal Effects Using a Single Proxy Variable
Silvan Vollmer, Niklas Pfister, Sebastian Weichwald
TLDR
This paper introduces SPICE, a method for identifying causal effects using a single proxy for unobserved confounders, and proposes SPICE-Net for estimation.
Key contributions
- Proves causal effects are identifiable using a single proxy variable under the SPICE completeness assumption.
- Extends prior proxy-based identifiability results to higher dimensions and more flexible relationships.
- Develops SPICE-Net, a neural network framework for estimating causal effects with discrete/continuous treatments.
Why it matters
Unobserved confounding hinders accurate causal inference. This work provides a theoretical foundation (SPICE) and a practical tool (SPICE-Net) to overcome this challenge using readily available proxy data. It significantly broadens the applicability of proxy-based causal identification methods.
Original 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.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.