Time Series Causal Discovery via Context-Conditioned and Causality-Augmented Pretraining

2026-05-26

AI Summary

Existing time series causal discovery methods exhibit limited generalization capabilities, particularly when applied to out-of-distribution (OOD) tasks involving varying causal mechanisms. To address this challenge, this work proposes PTCD, a pretraining framework for causal discovery that models dynamic causal structures through a dual-scale iterative attention mechanism and a context-routing Gaussian mixture model. By integrating interventional learning with a causal aliasing strategy during synthetic data pretraining, PTCD substantially enhances OOD generalization for both cross-task causal discovery and root cause identification. Empirical evaluations demonstrate that the proposed method achieves state-of-the-art performance across multiple real-world datasets.

Abstract

Causal discovery from time series is critical for many real-world applications, such as tracing the root causes of anomalies. Existing approaches typically rely on dataset-specific optimization, making it difficult to transfer their causal discovery capabilities to new time series governed by diverse causal mechanisms. In this paper, we propose \textbf{PTCD}, a novel \textbf{P}retraining framework for \textbf{T}ime-series \textbf{C}ausal \textbf{D}iscovery, which improves cross-task generalization through context-conditioned modeling and transferable causal augmentation. To model complex temporal causal dependencies, PTCD employs a dual-scale iterative attention mechanism to capture window-level causal relationships, and a Gaussian mixture with a context-level routing mechanism to handle heterogeneous exogenous distributions. To further address distribution shifts across causal graphs, PTCD adopts a pretraining paradigm on synthetic datasets that integrates intervention-based learning and a causal mixup strategy, promoting stable causal discovery and stronger generalization. Extensive experiments on multiple real-world out-of-distribution (OOD) datasets demonstrate that PTCD excels in both causal discovery and root cause identification.

Problem

• time series
• causal discovery
• cross-task generalization
• distribution shift
• out-of-distribution

Innovation

• causal discovery
• time series
• pretraining
• context-conditioned modeling
• causal augmentation