Identification of Nonparametric Dynamic Causal Model and Latent Process for Climate Analysis
Keywords: causal discovery; causal representation learning; climate analysis
Abstract: The study of learning causal structure with latent variables has advanced the understanding of the world by uncovering causal relationships and latent factors. However, in real-world scenarios, such as those in climate systems, causal relationships are often nonparametric, dynamic, and exist among both observed variables and latent variables. These challenges motivate us to consider a general setting in which causal relations are nonparametric and unrestricted in their occurrence, which is unconventional to current methods. To solve this problem, with the aid of 3-measurement in temporal structure, we theoretically show that both latent variables and processes can be identified up to minor indeterminacy under mild assumptions. Furthermore, we establish that the observed causal structure is identifiable if there is generation variability, roughly speaking, the latent variables induce sufficient variations in generating the noise terms, by the established functional equivalence. The primary idea of this framework is to learn causal representations from causally-related observations, and subsequently address this problem as a task of general nonlinear causal discovery. Based on these theoretical insights, we develop an estimation approach simultaneously learning both the observed causal structure, latent representation, and latent Markov network. Experimental results in simulation studies validate the theoretical foundations and demonstrate the effectiveness of the proposed methodology. In the climate data experiments, we show that it offers a powerful and in-depth understanding of the climate system.
Primary Area: causal reasoning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 7172
Loading