[论文] Twice Upon a Time: Complementary Representations for Time-Series Reaso...

论文概要

研究领域: NLP
作者: Hedi Zisling, Liran Nochumsohn, Ilan Naiman, Mihir Parmar, Omri Azencot
发布时间: 2026-10-15
arXiv: 2610.22327

中文摘要

时间序列语言模型（TSLM）实现对时间数据的语言理解和推理，但一个基本问题仍然存在：其时间序列表示应保留什么信息，应分解出什么变化？现有 TSLM 通常仅暴露单一表示，跨下游任务施加固定的归纳偏置。我们引入互补表示作为时间序列推理的设计原则。我们的模型结合两种视觉视图：线图保留绝对测量和观测坐标系，而归一化延迟嵌入分解正幅度重标定和偏移，同时暴露时间结构。基于 Qwen3-VL-8B，我们的双视图模型将 TSRBench 整体准确率从仅使用线图的 44.5% 提高到 53.2%，大幅超过最强的专用 TSLM 基线（36.7%）。互补表示为通用 TSLM 提供了有效的设计原则。

原文摘要

Time Series Language Models (TSLMs) enable language-based understanding and reasoning over temporal data, yet a fundamental question remains: what information should their time-series representations preserve, and what variation should they factor out? Existing TSLMs typically expose the language model to a single representation, imposing a fixed inductive bias across downstream tasks. We instead introduce complementary representations as a design principle for time-series reasoning. Our model combines two visual views: line plots preserve absolute measurements and the observed coordinate frame, while normalized delay embeddings factor out positive amplitude rescaling and offsets while exposing temporal structure. Dedicated visual encoders integrate both views into a shared language model ...

自动采集于 2026-10-17

#论文 #arXiv #NLP #小凯