用时间蒸馏让大模型学会预测时间序列，效果优于现有方法。
T-LLM: Teaching Large Language Models to Forecast Time Series via Temporal Distillation
- 通过轻量级教师模型传授预测行为，训练时注入时间结构监督。
- 在全量、少样本、零样本场景下均超越现有LLM方法，提升显著。
- 无需推理时依赖教师模型，部署简单高效，适合实际应用。
时间序列预测在诸多现实应用中至关重要。与视觉和语言数据不同，时间序列数据本质上随真实时间演进而积累，无法通过规模驱动的预训练单独解决。这一时间约束使得大语言模型（LLMs）获取预测能力面临挑战，现有方法多依赖表示对齐或推理时的时间模块，而非显式教授预测行为。我们提出T-LLM，一种时间蒸馏框架，通过在训练期间将轻量级时间教师模型的预测行为迁移给通用大模型，赋予其时间序列预测能力。教师模型结合趋势建模与频域分析提供结构化时间监督，推理时完全移除，仅由大模型独立完成预测。在基准数据集及传染病预测任务上的实验表明，T-LLM在全量、少样本、零样本设置下均持续优于现有基于LLM的预测方法，同时支持简单高效的部署流程。
原文摘要 · Abstract (English)
Time series forecasting plays a critical role in decision-making across many real-world applications. Unlike data in vision and language domains, time series data is inherently tied to the evolution of underlying processes and can only accumulate as real-world time progresses, limiting the effectiveness of scale-driven pretraining alone. This time-bound constraint poses a challenge for enabling large language models (LLMs) to acquire forecasting capability, as existing approaches primarily rely on representation-level alignment or inference-time temporal modules rather than explicitly teaching forecasting behavior to the LLM. We propose T-LLM, a temporal distillation framework that equips general-purpose LLMs with time series forecasting capability by transferring predictive behavior from a lightweight temporal teacher during training. The teacher combines trend modeling and frequency-domain analysis to provide structured temporal supervision, and is removed entirely at inference, leaving the LLM as the sole forecasting model. Experiments on benchmark datasets and infectious disease forecasting tasks demonstrate that T-LLM consistently outperforms existing LLM-based forecasting methods under full-shot, few-shot, and zero-shot settings, while enabling a simple and efficient deployment pipeline.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。