CTRL: Cross-Modal Transformer With Reinforcement Learning for Climate Causal Discovery

Abstract

Understanding the complex internal relationships of climate elements and revealing causal relationships in climate phenomena has attracted attention in the field of climate science and data mining. However, current causal discovery methods in climate science primarily rely on single-modal data, which limits the available information and complicates the accurate identification of climate causal relationships. In this article, we propose a novel climate causal discovery method based on cross-modal Transformer and reinforcement learning (RL), called CTRL. The method consists of three modules: unimodal feature representation, cross-modal feature fusion, and RL for causal discovery. Specifically, we first design a novel unimodal feature representation method that uses the seasonal-trend decomposition based on LOESS (STL) strategy to decompose air quality and meteorological features into trend and seasonal components, with the aim of effectively capturing temporal data patterns. Next, we develop a cross-modal feature fusion method using cross-attention to capture cross-modal interactions and adaptively fuse bimodal features. Finally, we employ RL to search for the optimal climate causal network with the best Bayesian information criterion (BIC) score. Numerous experiments show that CTRL outperforms single-modal approaches in climate causal discovery, achieving even better results when predicting a target climate variable using only the causal features identified by CTRL.