aiqbench analysis
This research demonstrates that in environmental causal inference, predictive accuracy of learned transport models does not guarantee reliable spillover estimates when exposure mappings are uncertain. The study highlights a critical gap where modern operator-learning approaches diverge from mechanistic models in estimating intervention effects, particularly for localized point-source scenarios. For AI/quantum practitioners, this underscores the necessity of validating causal mechanisms beyond mere loss minimization when deploying learned models in complex physical systems.
Causal Inference under Interference with Learned Exposure Mappings
arXiv:2608.19224v1 Announce Type: cross Abstract: Exposure mappings are often assumed to be known in causal spillover analyses. In environmental settings, however, they are typically induced by transport processes that are not directly observed and must instead be learned from pollution data. We study how uncertainty in learned transport processes propagates into exposure mappings and downstream spillover inference under interference. We compare mechanistic transport models with modern operator-learning approaches, including PDE, PINO, FNO, and GeoPT, using both simulation studies and an empirical analysis of California PM$_{2.5}$ data. In simulations, all four transport models achieved nearly identical pollution prediction accuracy, yet estimated spillover effects ranged from 1.78 to 2.27. Models that more accurately recovered the induced exposure mapping also produced spillover estimates closer to the true effect. Disagreement was modest for regional interventions but substantially larger for localized point-source interventions. The California analysis showed the same pattern: competing transport models produced similar predictions of observed PM${2.5}$ concentrations while implying different spillover effects under hypothetical pollution-control interventions. Our findings suggest that predictive agreement alone is insufficient for reliable causal inference when exposure mappings are learned rather than directly observed.
Abstract:Exposure mappings are often assumed to be known in causal spillover analyses. In environmental settings, however, they are typically induced by transport processes that are not directly observed and must instead be learned from pollution data. We study how uncertainty in learned transport processes propagates into exposure mappings and downstream spillover inference under interference. We compare mechanistic transport models with modern operator-learning approaches, including PDE, PINO, FNO, and GeoPT, using both simulation studies and an empirical analysis of California PM$_{2.5}$ data. In simulations, all four transport models achieved nearly identical pollution prediction accuracy, yet estimated spillover effects ranged from 1.78 to 2.27. Models that more accurately recovered the induced exposure mapping also produced spillover estimates closer to the true effect. Disagreement was modest for regional interventions but substantially larger for localized point-source interventions. The California analysis showed the same pattern: competing transport models produced similar predictions of observed PM${2.5}$ concentrations while implying different spillover effects under hypothetical pollution-control interventions. Our findings suggest that predictive agreement alone is insufficient for reliable causal inference when exposure mappings are learned rather than directly observed.
Methodology (stat.ME); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Machine Learning (cs.LG); Applications (stat.AP)
arXiv:2608.19224 [stat.ME]
(or arXiv:2608.19224v1 [stat.ME] for this version)
https://doi.org/10.48550/arXiv.2608.19224 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 · 284 words
Related coverage
Estimating Treatment Effects in Networks under Unknown Exposure Mappings
arXiv · Aug 4, 2026
Game Design Creativity evals
LessWrong · Oct 12, 2026
GIST-CMTF: Goal-State Inference for Causal Minimal Tool Filtering in LLM Agents
arXiv · Jun 16, 2026
What's Continual Learning, and Why Might We Expect To See It In Advanced LLM Agents?
LessWrong · Jun 12, 2026