Every Client Is an Environment: Federated De-confounding for Spatio-Temporal Forecasting

2026-07-27

AI Summary

This work addresses the limitations of existing federated spatiotemporal forecasting methods, which treat client heterogeneity as an optimization obstacle and struggle to adapt to environmental shifts. Instead, it innovatively regards each client as an independent causal environment, leveraging heterogeneity as evidence of environmental diversity. The paper proposes the first federated deconfounding framework grounded in a causal environment perspective: by learning a shared global prototype codebook, it extracts environment-invariant mechanisms to enable deconfounded spatiotemporal prediction. Theoretical analysis provides error bounds for deconfounding, and experiments demonstrate that the method significantly outperforms current federated approaches across multiple benchmarks, yielding transferable, interpretable, and communication-efficient environmental representations.

Abstract

Federated learning has emerged as a promising paradigm for spatio-temporal forecasting (STF), enabling collaborative model training without sharing raw observations. Existing federated STF methods primarily regard cross-client heterogeneity as an optimization challenge and mitigate it through personalized approaches. However, such heterogeneity fundamentally stems from diverse \emph{environmental conditions}, and these methods capture environment-specific forecasting patterns, hardly generalizing under environmental shifts. Our key insight is that the environmental diversity across federated clients should be exploited, as they provide \emph{complementary observations of the same underlying spatio-temporal system}. Based on this insight, we propose \method, a novel federated de-confounding framework that \textbf{treats clients as distinct causal environments}. \method leverages the client heterogeneity as distributed environmental evidence and learns a global prototype codebook to capture shared environmental regimes. We further derive a theoretical federated de-confounding bound that is linearly controlled by the averaged confounding strength. Extensive experiments demonstrate that \method consistently outperforms federated baselines, while providing transferable, interpretable, and communication-efficient environmental representations.

Problem

• federated learning
• spatio-temporal forecasting
• environmental heterogeneity
• confounding
• generalization

Innovation

• federated de-confounding
• causal environments
• spatio-temporal forecasting
• environmental heterogeneity
• prototype codebook