Adapting LLMs to Time Series Forecasting via Temporal Heterogeneity Modeling and Representation Alignment

Recent advances have demonstrated that Large Language Models (LLMs) can be effectively adapted for time series forecasting, revealing strong potential beyond natural language tasks. However, their performance remains constrained by two fundamental challenges: the inherent heterogeneity of temporal patterns and the modality gap between continuous numerical signals and discrete language representations. In this work, we propose TALON (Temporal-heterogeneity And Language-Oriented Network), a unified framework that enhances LLM-based forecasting by modeling temporal heterogeneity and promoting representation alignment. Specifically, we design a Heterogeneous Temporal Encoder that partitions multivariate time series into structurally coherent segments, enabling localized expert modeling across diverse temporal patterns. To bridge the modality gap, we introduce a Representation Alignment Module that projects temporal features toward LLM-compatible representations, making time series more amenable to LLMs and thereby unlocking their modeling potential, while eliminating the need for handcrafted prompts during inference. Extensive experiments on seven real-world benchmarks demonstrate that TALON achieves superior performance across all datasets, with average MSE improvements of up to 11% over recent state-of-the-art methods, while maintaining higher efficiency. These results underscore the effectiveness of incorporating both pattern-aware modeling and representation-level alignment when adapting LLMs for time series forecasting. The code is available at: https://github.com/syrGitHub/TALON.