GST-UNet: Spatiotemporal Causal Inference with Time-Varying Confounders

2025-02-07

AI Summary

Addressing the challenge of causal effect estimation in observational spatiotemporal data—particularly feedback bias induced by time-varying confounders—this paper proposes the first end-to-end framework integrating iterative G-computation with a spatiotemporal U-Net. Departing from conventional reliance on strong structural assumptions (e.g., no interference, static confounding), our approach enhances spatial modeling via graph convolution and employs regression-based iterative G-computation to infer counterfactual potential outcomes, explicitly capturing nonlinear spatiotemporal dynamics and feedback loops among confounders, interventions, and outcomes. Evaluated on both synthetic data and real-world 2018 California wildfire data, the method achieves significantly improved accuracy in causal effect estimation. Notably, it provides the first precise quantification of the time-varying causal impact of wildfire smoke on respiratory hospitalization rates.

Technology Category

• Reasoning under Uncertainty: Causality
• Machine Learning: Causal Learning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Estimating causal effects from spatiotemporal data is a key challenge in fields such as public health, social policy, and environmental science, where controlled experiments are often infeasible. However, existing causal inference methods relying on observational data face significant limitations: they depend on strong structural assumptions to address spatiotemporal challenges $unicode{x2013}$ such as interference, spatial confounding, and temporal carryover effects $unicode{x2013}$ or fail to account for $ extit{time-varying confounders}$. These confounders, influenced by past treatments and outcomes, can themselves shape future treatments and outcomes, creating feedback loops that complicate traditional adjustment strategies. To address these challenges, we introduce the $ extbf{GST-UNet}$ ($ extbf{G}$-computation $ extbf{S}$patio-$ extbf{T}$emporal $ extbf{UNet}$), a novel end-to-end neural network framework designed to estimate treatment effects in complex spatial and temporal settings. The GST-UNet leverages regression-based iterative G-computation to explicitly adjust for time-varying confounders, providing valid estimates of potential outcomes and treatment effects. To the best of our knowledge, the GST-UNet is the first neural model to account for complex, non-linear dynamics and time-varying confounders in spatiotemporal interventions. We demonstrate the effectiveness of the GST-UNet through extensive simulation studies and showcase its practical utility with a real-world analysis of the impact of wildfire smoke on respiratory hospitalizations during the 2018 California Camp Fire. Our results highlight the potential of GST-UNet to advance spatiotemporal causal inference across a wide range of policy-driven and scientific applications.

Problem

• Estimating causal effects from spatiotemporal data.
• Addressing time-varying confounders in causal inference.
• Handling complex spatial and temporal causal dynamics.

Innovation

• GST-UNet neural network
• time-varying confounders adjustment
• spatiotemporal causal inference