InvarGC: Invariant Granger Causality for Heterogeneous Interventional Time Series under Latent Confounding

2025-10-21

AI Summary

Traditional Granger causality tests rely on linearity, causal sufficiency (i.e., no latent confounders), and known intervention targets, limiting their applicability to nonlinear, heterogeneous real-world settings. To address this, we propose InvarGC—a novel method that identifies interventions and discriminates environments at the edge level without assuming causal completeness or prior knowledge of interventions. InvarGC integrates nonlinear Granger modeling with cross-environment invariance learning, leveraging deep temporal models to capture complex dynamics and introducing an environment-label inference mechanism to mitigate unobserved confounding. We provide theoretical guarantees for identifiability of the underlying causal structure. Extensive experiments on synthetic and real-world datasets demonstrate that InvarGC significantly outperforms state-of-the-art methods, effectively suppressing spurious correlations and enhancing robustness and interpretability of causal discovery under heterogeneous interventions.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

Granger causality is widely used for causal structure discovery in complex systems from multivariate time series data. Traditional Granger causality tests based on linear models often fail to detect even mild non-linear causal relationships. Therefore, numerous recent studies have investigated non-linear Granger causality methods, achieving improved performance. However, these methods often rely on two key assumptions: causal sufficiency and known interventional targets. Causal sufficiency assumes the absence of latent confounders, yet their presence can introduce spurious correlations. Moreover, real-world time series data usually come from heterogeneous environments, without prior knowledge of interventions. Therefore, in practice, it is difficult to distinguish intervened environments from non-intervened ones, and even harder to identify which variables or timesteps are affected. To address these challenges, we propose Invariant Granger Causality (InvarGC), which leverages cross-environment heterogeneity to mitigate the effects of latent confounding and to distinguish intervened from non-intervened environments with edge-level granularity, thereby recovering invariant causal relations. In addition, we establish the identifiability under these conditions. Extensive experiments on both synthetic and real-world datasets demonstrate the competitive performance of our approach compared to state-of-the-art methods.

Problem

• Detecting nonlinear causal relationships in time series
• Addressing latent confounding effects in causal discovery
• Identifying interventions in heterogeneous environments without prior knowledge

Innovation

• Leverages cross-environment heterogeneity to mitigate latent confounding
• Distinguishes intervened from non-intervened environments with edge-level granularity
• Recovers invariant causal relations under heterogeneous interventional time series