Time-series diagnostic reasoning is essential for many applications, yet existing solutions face a persistent gap: general reasoning large language models (GRLMs) possess strong reasoning skills but lack the domain-specific knowledge to understand complex time-series patterns. Conversely, fine-tuned time-series LLMs (TSLMs) understand these patterns but lack the capacity to generalize reasoning for more complicated questions. To bridge this gap, we propose a hybrid knowledge-injection framework that injects TSLM-generated insights directly into GRLM's reasoning trace, thereby achieving strong time-series reasoning with in-domain knowledge. As collecting data for knowledge injection fine-tuning is costly, we further leverage a reinforcement learning-based approach with verifiable rewards (RLVR) to elicit knowledge-rich traces without human supervision, then transfer such an in-domain thinking trace into GRLM for efficient knowledge injection. We further release SenTSR-Bench, a multivariate time-series-based diagnostic reasoning benchmark collected from real-world industrial operations. Across SenTSR-Bench and other public datasets, our method consistently surpasses TSLMs by 9.1%-26.1% and GRLMs by 7.9%-22.4%, delivering robust, context-aware time-series diagnostic insights.
@article{arxiv.2602.19455,
title = {SenTSR-Bench: Thinking with Injected Knowledge for Time-Series Reasoning},
author = {Zelin He and Boran Han and Xiyuan Zhang and Shuai Zhang and Haotian Lin and Qi Zhu and Haoyang Fang and Danielle C. Maddix and Abdul Fatir Ansari and Akash Chandrayan and Abhinav Pradhan and Bernie Wang and Matthew Reimherr},
journal= {arXiv preprint arXiv:2602.19455},
year = {2026}
}
Comments
Accepted by the 29th International Conference on Artificial Intelligence and Statistics (AISTATS 2026)