arXiv 2605.034602026-05cs.AIcs.LG

FinSTaR：面向时间序列推理模型的金融推理

FinSTaR: Towards Financial Reasoning with Time Series Reasoning Models

LG AI Research（LG人工智能研究）

Seunghan Lee, Jun Seo, Jaehoon Lee, Sungdong Yoo, Minjae Kim, Tae Yoon Lim, Dongwan Kang, Hwanil Choi, Soonyoung Lee, Wonbin Ahn

备注：: EMNLP Industry track 2026, KDD Workshop on SciSoc Agents & LLMs 2026 (Oral Presentation)

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对时间序列推理模型在金融领域的失效问题，提出基于2x2能力分类法的FinSTaR模型，通过Compute-in-CoT和Scenario-Aware CoT策略在FinTSR-Bench基准上达到78.9%平均准确率。

摘要：

时间序列推理模型在通用领域表现出色，但在具有独特特征的金融领域却持续失败。我们提出一个通用的2x2能力分类法，通过交叉1)单实体与多实体分析，以及2)当前状态评估与未来行为预测来划分TSRM能力。我们在金融领域实例化该分类法——其中确定性评估与随机性预测的区分尤为关键——形成十个金融推理任务，并基于标普股票构建FinTSR-Bench基准。为此，我们提出FinSTaR（金融时间序列思考与推理），在FinTSR-Bench上训练，并针对每个类别采用不同的思维链策略。对于评估（确定性，即可从可观测数据计算得出），我们采用Compute-in-CoT，一种程序化思维链，使模型能够直接从原始价格推导答案。对于预测（本质上是随机的，即受不可观测因素影响），我们采用场景感知思维链，在做出判断前生成多种场景，模拟金融分析师在不确定性下的推理方式。所提方法在FinTSR-Bench上达到78.9%的平均准确率，显著优于LLM和TSRM基线。此外，我们展示了四个能力类别通过联合训练具有互补性和相互增强性，并且场景感知思维链相比标准思维链持续提升预测准确率。代码已公开：https://github.com/seunghan96/FinSTaR。

英文摘要：

Time series (TS) reasoning models (TSRMs) have shown promising capabilities in general domains, yet they consistently fail on financial domain, which exhibit unique characteristics. We propose a general 2 x 2 capability taxonomy for TSRMs by crossing 1) single-entity vs. multi-entity analysis with 2) assessment of the current state vs. prediction of future behavior. We instantiate this taxonomy in the financial domain---where the distinction between deterministic assessment and stochastic prediction is particularly critical---as ten financial reasoning tasks, forming the FinTSRBench benchmark based on S&P stocks. To this end, we propose FinSTaR (Financial Time Series Thinking and Reasoning), trained on FinTSR-Bench with distinct chain-of-thought (CoT) strategies tailored to each category. For assessment, which is deterministic, we employ Compute-in-CoT, a programmatic CoT that enables models to derive answers directly from raw prices. For prediction, which is inherently stochastic, we adopt Scenario-Aware CoT, which generates diverse scenarios before making a judgment, mirroring how financial analysts reason under uncertainty. FinSTaR achieves 78.9% average accuracy on FinTSRBench, substantially outperforming LLM and TSRM baselines. Furthermore, we show that the four capability categories are complementary and mutually reinforcing through joint training. Code is publicly available at: https://github.com/seunghan96/FinSTaR.

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。

AI总结：

摘要：

英文摘要：

相关论文