Correlation to Causation: A Causal Deep Learning Framework for Arctic Sea Ice Prediction
This work addresses the challenge of improving robustness and interpretability in predictive modeling for dynamic, high-dimensional systems like Arctic sea ice, with incremental contributions by integrating existing causal discovery methods into a deep learning framework.
The paper tackled the problem of spurious correlations limiting robustness in Arctic sea ice prediction by proposing a causality-driven deep learning framework, resulting in enhanced predictive accuracy and interpretability across multiple lead times.
Traditional machine learning and deep learning techniques rely on correlation-based learning, often failing to distinguish spurious associations from true causal relationships, which limits robustness, interpretability, and generalizability. To address these challenges, we propose a causality-driven deep learning framework that integrates Multivariate Granger Causality (MVGC) and PCMCI+ causal discovery algorithms with a hybrid deep learning architecture. Using 43 years (1979-2021) of daily and monthly Arctic Sea Ice Extent (SIE) and ocean-atmospheric datasets, our approach identifies causally significant factors, prioritizes features with direct influence, reduces feature overhead, and improves computational efficiency. Experiments demonstrate that integrating causal features enhances the deep learning model's predictive accuracy and interpretability across multiple lead times. Beyond SIE prediction, the proposed framework offers a scalable solution for dynamic, high-dimensional systems, advancing both theoretical understanding and practical applications in predictive modeling.