![[2602.22532] Coarse-to-Fine Learning of Dynamic Causal Structures](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.22532] Coarse-to-Fine Learning of Dynamic Causal Structures
Summary
The paper presents DyCausal, a framework for learning dynamic causal structures in time series data, addressing challenges of time-varying relationships through a coarse-to-fine approach.
Why It Matters
Understanding dynamic causal structures is crucial for modeling complex systems in fields like economics and neuroscience. DyCausal offers a novel method that enhances causal discovery, which can lead to better predictions and insights in various applications.
Key Takeaways
- DyCausal framework captures dynamic causal structures effectively.
- Utilizes convolutional networks for initial coarse-grained analysis.
- Implements linear interpolation for fine-grained causal graph recovery.
- Acyclic constraints improve efficiency and stability in causal discovery.
- Demonstrated superior performance on synthetic and real-world datasets.
Computer Science > Machine Learning arXiv:2602.22532 (cs) [Submitted on 26 Feb 2026] Title:Coarse-to-Fine Learning of Dynamic Causal Structures Authors:Dezhi Yang, Qiaoyu Tan, Carlotta Domeniconi, Jun Wang, Lizhen Cui, Guoxian Yu View a PDF of the paper titled Coarse-to-Fine Learning of Dynamic Causal Structures, by Dezhi Yang and 5 other authors View PDF HTML (experimental) Abstract:Learning the dynamic causal structure of time series is a challenging problem. Most existing approaches rely on distributional or structural invariance to uncover underlying causal dynamics, assuming stationary or partially stationary causality. However, these assumptions often conflict with the complex, time-varying causal relationships observed in real-world systems. This motivates the need for methods that address fully dynamic causality, where both instantaneous and lagged dependencies evolve over time. Such a setting poses significant challenges for the efficiency and stability of causal discovery. To address these challenges, we introduce DyCausal, a dynamic causal structure learning framework. DyCausal leverages convolutional networks to capture causal patterns within coarse-grained time windows, and then applies linear interpolation to refine causal structures at each time step, thereby recovering fine-grained and time-varying causal graphs. In addition, we propose an acyclic constraint based on matrix norm scaling, which improves efficiency while effectively constraining loops in evol...
