Twice Upon a Time: Complementary Representations for Time-Series Reasoning
arXiv cs.CL · 16 Oct 2026, 03:52 AEST
A new model for time-series reasoning introduces complementary representations to enhance understanding of temporal data. This approach combines line plots and normalized delay embeddings, leading to significant improvements in prediction and reasoning accuracy.
Key facts
- 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?
- 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.
- Built on Qwen3-VL-8B, our dual-view model improves overall TSRBench accuracy from 44.5% with line plots alone to 53.2%, substantially exceeding the strongest evaluated dedicated TSLM baseline at 36.7%.
- Together, these results demonstrate that complementary representations provide an effective design principle for general-purpose TSLMs, enabling models to exploit temporal structure without sacrificing access to the original measurement frame.
Read aloud
A new model for time-series reasoning introduces complementary representations to enhance understanding of temporal data. This approach combines line plots and normalized delay embeddings, leading to significant improvements in prediction and reasoning accuracy. 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? 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. Built on Qwen3-VL-8B, our dual-view model improves overall TSRBench accuracy from forty four point five percent with line plots alone to fifty three point two percent, substantially exceeding the strongest evaluated dedicated TSLM baseline at thirty six point seven percent. Together, these results demonstrate that complementary representations provide an effective design principle for general-purpose TSLMs, enabling models to exploit temporal structure without sacrificing access to the original measurement frame.
Sources in this cluster
- Twice Upon a Time: Complementary Representations for Time-Series Reasoning
arXiv cs.CL · 16 Oct 2026, 03:52 AEST