aiqbench analysis
This paper introduces PO-Flow, a continuous normalizing flow framework designed to generate potential outcomes and counterfactuals for individualized decision-making in causal inference. By integrating cross-fitted inverse-probability weighting into the flow-matching loss, the method effectively addresses treatment imbalance while recovering individual counterfactuals under specific structural assumptions. The approach demonstrates superior performance on both scalar benchmarks and high-dimensional multivariate settings compared to existing baselines.
Generative Flows for Potential Outcomes and Counterfactuals
arXiv:2505.16051v5 Announce Type: replace-cross Abstract: Average treatment effects summarize how interventions change outcomes, but individualized decisions require predictions and outcome distributions of what would have happened to the same individual under an alternative treatment. We develop PO-Flow, a continuous normalizing flow framework that represents treatment-specific outcome distributions through invertible transport maps. It supports potential-outcome generation, conditional average treatment-effect estimation, factual-conditioned counterfactual prediction, and density evaluation. To address treatment imbalance in observational data, we incorporate cross-fitted inverse-probability weighting into the flow-matching loss. We establish sufficient conditions under which encoding a factual outcome and decoding under an alternative treatment recovers the individual counterfactual. Under shared exogenous noise and a common encoded reference distribution, scalar recovery follows from increasing noise-to-latent maps. For multivariate outcomes, we assume triangularity or a convex-gradient structure. These conditions distinguish fitting outcome distributions from recovering their causal correspondence. Experiments on four scalar benchmarks and seven multivariate or image settings, with dimensions up to 784, compare against ten scalar and five multivariate baselines. Controlled ablations show IPW improves all reported scalar counterfactual RMSE, PEHE, and potential-outcome RMSE. PO-Flow achieves the lowest PEHE and potential-outcome RMSE on all four scalar benchmarks and attains or ties the lowest mean in 26 of 35 multivariate dataset--metric comparisons.
Abstract:Average treatment effects summarize how interventions change outcomes, but individualized decisions require predictions and outcome distributions of what would have happened to the same individual under an alternative treatment. We develop PO-Flow, a continuous normalizing flow framework that represents treatment-specific outcome distributions through invertible transport maps. It supports potential-outcome generation, conditional average treatment-effect estimation, factual-conditioned counterfactual prediction, and density evaluation. To address treatment imbalance in observational data, we incorporate cross-fitted inverse-probability weighting into the flow-matching loss. We establish sufficient conditions under which encoding a factual outcome and decoding under an alternative treatment recovers the individual counterfactual. Under shared exogenous noise and a common encoded reference distribution, scalar recovery follows from increasing noise-to-latent maps. For multivariate outcomes, we assume triangularity or a convex-gradient structure. These conditions distinguish fitting outcome distributions from recovering their causal correspondence. Experiments on four scalar benchmarks and seven multivariate or image settings, with dimensions up to 784, compare against ten scalar and five multivariate baselines. Controlled ablations show IPW improves all reported scalar counterfactual RMSE, PEHE, and potential-outcome RMSE. PO-Flow achieves the lowest PEHE and potential-outcome RMSE on all four scalar benchmarks and attains or ties the lowest mean in 26 of 35 multivariate dataset--metric comparisons.
Machine Learning (stat.ML); Machine Learning (cs.LG)
arXiv:2505.16051 [stat.ML]
(or arXiv:2505.16051v5 [stat.ML] for this version)
https://doi.org/10.48550/arXiv.2505.16051 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 294 words
Related coverage
CoCoGEC: Counterfactual Generation for Robust Grammatical Error Correction
arXiv · Jun 17, 2026
ConTex: Reformulating Counterfactual Generation For Time Series Forecasting
arXiv · Jun 17, 2026
Class-Aware Reinforcement Learning for Counterfactual Explanation Generation
arXiv · Jul 31, 2026
iFlip: Iterative Feedback-driven Counterfactual Example Refinement
arXiv · Aug 27, 2026