aiqbench analysis
This research advances causal foundation models by establishing a canonical prior over structural causal models to bound counterfactuals under unobserved confounding. By translating partially-identifiable causal effect estimation into learning distributions over functions, the work provides a rigorous framework for deriving intervention effects from observational data where multiple values remain compatible with observations.
Foundation Models for Partial Causal Identification
arXiv:2608.20841v1 Announce Type: new Abstract: This paper investigates the development of causal foundation models for bounding the effect of interventions and counterfactuals from observational data. We show that a canonical prior can be defined with full support over the space of structural causal models with discrete observables. With this canonical prior, we translate the problem of bounding counterfactuals into that of learning distributions over functions that map data (and possibly structural assumptions) to a causal query of interest. This extends the promising causal foundational modelling paradigm to the estimation of partially-identifiable causal effects, i.e., under unobserved confounding, where multiple values are equally compatible with the observed data and prior structural assumptions.
Abstract:This paper investigates the development of causal foundation models for bounding the effect of interventions and counterfactuals from observational data. We show that a canonical prior can be defined with full support over the space of structural causal models with discrete observables. With this canonical prior, we translate the problem of bounding counterfactuals into that of learning distributions over functions that map data (and possibly structural assumptions) to a causal query of interest. This extends the promising causal foundational modelling paradigm to the estimation of partially-identifiable causal effects, i.e., under unobserved confounding, where multiple values are equally compatible with the observed data and prior structural assumptions.
Artificial Intelligence (cs.AI)
arXiv:2608.20841 [cs.AI]
(or arXiv:2608.20841v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.20841 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
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 · 200 words
Related coverage
What You Observe Determines How You Identify Causal Effects: Evaluating Causal Models across Observational Views
arXiv · Sep 30, 2026
Use What You Know: Causal Foundation Models with Partial Graphs
arXiv · Sep 30, 2026
PIRAMID: Progress and Plans
LessWrong · Jul 27, 2026
Causal DAG Identification for Count Data via Poisson Thinning Structural Equation Models
arXiv · Sep 9, 2026