SenTSR-Bench: 带注入知识的思考以实现时间序列推理
SenTSR-Bench: Thinking with Injected Knowledge for Time-Series Reasoning
AI总结：
SenTSR-Bench通过注入领域知识提升时间序列推理能力，结合强化学习方法实现高效的知识注入，方法在多个数据集上显著优于现有模型。
摘要：
时间序列诊断推理对于许多应用至关重要，但现有解决方案面临持续存在的差距：通用推理大语言模型（GRLMs）具有强大的推理能力，但缺乏领域特定的知识来理解复杂的时序模式。相反，微调的时间序列大语言模型（TSLMs）能够理解这些模式，但缺乏对更复杂问题进行泛化推理的能力。为弥合这一差距，我们提出了一种混合知识注入框架，将TSLM生成的洞察直接注入到GRLM的推理轨迹中，从而实现具有领域知识的时间序列推理。由于收集用于知识注入微调的数据成本高昂，我们进一步利用基于强化学习的方法，结合可验证的奖励（RLVR）来获取富含知识的轨迹，而无需人工监督，然后将此类领域内的思考轨迹转移到GRLM中以实现高效的的知识注入。我们进一步发布了SenTSR-Bench，这是一个基于多变量时间序列的诊断推理基准，收集自真实世界工业操作。在SenTSR-Bench和其他公开数据集上，我们的方法在TSLMs上始终领先9.1%-26.1%，在GRLMs上领先7.9%-22.4%，提供了稳健、上下文感知的时间序列诊断洞察。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
