Every Client Is an Environment: Federated De-confounding for Spatio-Temporal Forecasting
2026-07-27 • Machine Learning
Machine LearningArtificial Intelligence
AI summaryⓘ
The authors study a method for forecasting data that changes over time and space without sharing raw data between sources, called federated learning. They point out that differences between clients usually come from different environmental conditions, which existing methods don't handle well when environments change. Their new approach, method, treats each client as a unique environment and uses these differences to learn shared patterns across all clients. This helps improve forecasting accuracy and makes the learned models easier to understand and share efficiently. They also provide a theoretical explanation for why their approach reduces confounding effects caused by environmental differences.
federated learningspatio-temporal forecastingcross-client heterogeneityenvironmental conditionscausal environmentsde-confoundingglobal prototype codebookenvironmental regimesconfounding strengthcommunication efficiency
Authors
Qingxiang Liu, Anqi Liang, Heng Wang, Yuxuan Liang
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.