LLM-GC: Temporal-Semantic Disentanglement with Retrieval Augmentation to Activate LLM's Ability for Multimodal Granger Causal Discovery

Bo Liu, Hongyan Li, Shenda Hong

Keywords: Multimodel Granger Causal Discovery, Cross-Modality Alignment, Dual-Modal Time Series Encoding, Causality-Aware Self-Attention

Abstract: Recent advances in neural Granger causal methods have shown promise in modeling temporal nonlinear dependencies. However, existing approaches remain confined to raw time-series data, inherently lacking contextual semantics and tending to overfit, which undermines their real-world applicability. To address these challenges, we propose \textbf{LLM-GC}, a novel LLM-empowered multimodal Granger causality discovery framework that enriches unimodal temporal dynamics with semantic priors and world knowledge distilled from large language models (LLMs). LLM-GC leverages dual-modality encoding to capture and align both temporal and contextual dynamics by Cross-Modal Dual Retrieval while avoiding causal entanglement across modalities. To extract multimodal causal features, we introduce a causality-aware self-attention mechanism by simply inverting the conventional self-attention structure, enabling a shared causality augmenter to effectively highlight consistent causal patterns across modalities. LLM-GC is the first to bridge LLMs and Granger causality, and experiments on synthetic and real-world benchmark datasets demonstrate that LLM-GC outperforms existing state-of-the-art methods in Granger causal discovery.

Primary Area: causal reasoning

Submission Number: 1548