FreSia：面向多元时间序列分析的频率-语义实例化与对齐
FreSia: Frequency-Semantic Instantiation and Alignment for Multivariate Time Series Analysis
浏览论文内容
中文总结 AI 辅助
FreSia通过频率引导提示和全局上下文学习，将LLM语义空间与时间序列频域对齐，在八个基准上平均提升MSE 13.48%、MAE 8.06%。
中文摘要 AI 辅助
大型语言模型（LLMs）在多元时间序列预测和异常检测中展现出强大潜力。现有研究主要通过直接数值标记化或启发式文本描述将时间信息注入LLMs。然而，LLMs在感知数值时间序列的底层结构模式方面仍面临困难，尤其是被离散数值标记掩盖的季节性和趋势成分。为弥合这一差距，我们提出FreSia，一个频率感知框架，在LLMs的语义空间与时间序列的频率空间之间建立有效对齐。具体而言，FreSia中的FGPrompt（频率引导提示机制）提炼时间序列的频域结构，并将其投影到针对LLMs语义空间定制的提示中。此外，我们引入全局驱动上下文学习（GCL）组件，利用全局CLS驱动探针生成全局上下文，以弥合时频域差距并融合多模态信息。在八个预测基准上的实验表明，FreSia在MSE和MAE上分别实现了13.48%和8.06%的平均改进。
英文摘要
Large Language Models (LLMs) have shown strong potential in multivariate time series forecasting and anomaly detection. Existing studies predominantly inject temporal information into LLMs via direct numerical tokenization or heuristic textual descriptions. However, LLMs still face difficulty in perceiving the underlying structural patterns of numerical time series, particularly the seasonal and trend components obscured by discrete numerical tokens. To bridge this gap, we propose FreSia, a frequency-aware framework that establishes an effective alignment between the semantic space of LLMs and the frequency space of time series. Specifically, FGPrompt, a Frequency-Guided Prompt mechanism within FreSia, distills the frequency-domain structures of time series and projects them into prompts tailored to the semantic space of LLMs. Furthermore, we introduce a Global-driven Context Learning (GCL) component, which uses a global CLS-driven probe to generate global context to bridge the time-frequency domain gap and fuse the multi-modal information. Experiments on eight forecasting benchmarks show that FreSia achieves average improvements of 13.48% and 8.06% in MSE and MAE, respectively.
发表机构
- Beijing Institute of Technology（北京理工大学）
- Singapore Management University（新加坡管理大学）
机构由 AI 辅助整理，请以论文原文为准。
