Mitigating Unobserved Confounding via Diffusion Probabilistic Models

ICLR 2025 Conference Submission13081 Authors

Keywords: causal inference

Abstract: Learning Conditional average treatment effect estimation from observational data is a challenging task due to the existence of unobserved confounders. Previous methods mostly focus on assuming the Ignorability assumption ignoring the unobserved confounders or overlooking the impact of an a priori knowledge on the generation process of the latent variable, which can be quite impractical in real-world scenarios. Motivated by the recent advances in the latent variable modeling, we propose to capture the unobserved latent space using diffusion model, and accordingly to estimate the causal effect. More concretely, we build on the reverse diffusion process for the unobserved confounders as a Markov chain conditioned on an apriori knowledge. In order to implement our model in a feasible way, we derive the variational bound in closed form. In the experiments, we compare our model with the state-of-the-art methods based on both synthetic and real-world datasets, demonstrating consistent improvements of our model.

Primary Area: causal reasoning

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 13081