Select, then Balance: A Plug-and-Play Framework for Exogenous-Aware Spatio-Temporal Forecasting

2025-09-06

AI Summary

In spatiotemporal forecasting, modeling exogenous variables faces two key challenges: (1) heterogeneous impacts of different exogenous variables on the target system, and (2) imbalance between historical and future exogenous information. To address these, we propose a novel “select-then-balance” paradigm. First, we design a latent-space gated expert module to dynamically select salient external signals. Second, we construct a dual-branch spatiotemporal backbone network that separately captures temporal–spatial dependencies in historical and future exogenous variables, augmented with a context-aware weighting mechanism for dynamic balancing. To our knowledge, this is the first framework unifying gated selection and causal temporal balancing within a shared latent space, substantially enhancing model robustness and generalization. Extensive experiments on multiple real-world spatiotemporal datasets demonstrate superior performance over state-of-the-art methods in accuracy, generality, robustness, and computational efficiency.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Spatio-temporal forecasting aims to predict the future state of dynamic systems and plays an important role in multiple fields. However, existing solutions only focus on modeling using a limited number of observed target variables. In real-world scenarios, exogenous variables can be integrated into the model as additional input features and associated with the target signal to promote forecast accuracy. Although promising, this still encounters two challenges: the inconsistent effects of different exogenous variables to the target system, and the imbalance effects between historical variables and future variables. To address these challenges, this paper introduces model, a novel framework for modeling underline{exo}genous variables in underline{s}patio-underline{t}emporal forecasting, which follows a ``select, then balance'' paradigm. Specifically, we first construct a latent space gated expert module, where fused exogenous information is projected into a latent space to dynamically select and recompose salient signals via specialized sub-experts. Furthermore, we design a siamese network architecture in which recomposed representations of past and future exogenous variables are fed into dual-branch spatio-temporal backbones to capture dynamic patterns. The outputs are integrated through a context-aware weighting mechanism to achieve dynamic balance during the modeling process. Extensive experiments on real-world datasets demonstrate the effectiveness, generality, robustness, and efficiency of our proposed framework.

Problem

• Modeling inconsistent effects of exogenous variables
• Balancing historical and future variable impacts
• Integrating exogenous data for spatio-temporal forecasting

Innovation

• Latent space gated expert module for exogenous selection
• Siamese network architecture with dual-branch spatio-temporal backbones
• Context-aware weighting mechanism for dynamic balance integration