FiCoTS: Fine-to-Coarse LLM-Enhanced Hierarchical Cross-Modality Interaction for Time Series Forecasting

2025-11-28

AI Summary

To address the semantic gap between time-series and textual modalities—where large language models (LLMs) struggle to adapt effectively for time-series forecasting—this paper proposes FiCoTS, a novel framework that repositions the LLM as a “semantic enhancer” rather than a direct predictor. FiCoTS introduces a first-of-its-kind three-level cross-modal interaction paradigm—token-level → feature-level → decision-level—spanning fine-grained to coarse-grained abstraction. It employs dynamic heterogeneous graphs to model multi-source textual dependencies, global cross-attention for precise modality alignment, and a gated fusion network for progressive information integration, thereby bridging semantic discrepancies. Extensive experiments on seven real-world benchmark datasets demonstrate statistically significant improvements over state-of-the-art methods, validating FiCoTS’s superior accuracy, robustness, and generalization capability across diverse forecasting scenarios.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Time series forecasting is central to data analysis and web technologies. The recent success of Large Language Models (LLMs) offers significant potential for this field, especially from the cross-modality aspect. Most methods adopt an LLM-as-Predictor paradigm, using LLM as the forecasting backbone and designing modality alignment mechanisms to enable LLM to understand time series data. However, the semantic information in the two modalities of time series and text differs significantly, making it challenging for LLM to fully understand time series data. To mitigate this challenge, our work follows an LLM-as-Enhancer paradigm to fully utilize the advantage of LLM in text understanding, where LLM is only used to encode text modality to complement time series modality. Based on this paradigm, we propose FiCoTS, an LLM-enhanced fine-to-coarse framework for multimodal time series forecasting. Specifically, the framework facilitates progressive cross-modality interaction by three levels in a fine-to-coarse scheme: First, in the token-level modality alignment module, a dynamic heterogeneous graph is constructed to filter noise and align time series patches with text tokens; Second, in the feature-level modality interaction module, a global cross-attention mechanism is introduced to enable each time series variable to connect with relevant textual contexts; Third, in the decision-level modality fusion module, we design a gated network to adaptively fuse the results of the two modalities for robust predictions. These three modules work synergistically to let the two modalities interact comprehensively across three semantic levels, enabling textual information to effectively support temporal prediction. Extensive experiments on seven real-world benchmarks demonstrate that our model achieves state-of-the-art performance. The codes will be released publicly.

Problem

• Enhances time series forecasting using LLMs for cross-modality interaction
• Addresses semantic gap between time series and text modalities
• Proposes hierarchical framework for robust multimodal prediction integration

Innovation

• LLM-as-Enhancer paradigm for text encoding
• Fine-to-coarse hierarchical cross-modality interaction
• Dynamic graph, cross-attention, and gated fusion modules