arXiv：cs.LG· Haoxin Liu, Deqing Fu, Lucheng Fu, Yiqiao Jin, Rushi Qiang, Harshavardhan Kamarthi, B. Aditya Prakash·· 3 小时前
TIMERSI：面向时间序列基础模型的自改进智能体部署系统
TIMERSI: Towards Self-Improved Agentic Deployment for Time-Series Foundation Models
AI 导读
TIMERSI 是一个围绕冻结预测模型的自改进智能体部署系统，由 coding agent 将任务描述转化为可执行的 pipeline 修改，并在历史预测窗口上测试、依据误差迭代修订。在 TIME 和 FEVbench 的 198 个任务上，TIMERSI 在六项点预测与概率预测指标上超越所有提交结果。实验还验证了其在不同 coding agent 与预测模型上的泛化能力。
正文
Abstract:Time-series foundation models provide strong zero-shot forecast, but deployment for real- world applications still requires task-specific designs for model inputs and outputs. Direct inference can leave domain knowledge and re- curring forecast errors unaddressed, while tai- lored pipelines require substantial human ex- pertise and effort. We introduce TIMERSI, a self-improving agentic deployment system around frozen forecasting models. A cod- ing agent translates task descriptions into ex- ecutable pipeline changes. It tests each re- vision on historical forecasting windows and uses observed errors to guide subsequent re- visions. During deployment, an adaptive se- lector chooses among the resulting pipelines using completed forecasts under similar con- ditions. Across 198 tasks on TIME and FEV- bench, TIMERSI outperforms all submissions on six point and probabilistic forecasting met- rics. Controlled comparisons support the value of feedback-guided revision over non-iterative pipeline search. Further experiments verify the generalization with different coding agents and forecasting models.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.19725 [cs.LG] |
| (or arXiv:2610.19725v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.19725 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Haoxin Liu [view email]
[v1]
Wed, 14 Oct 2026 07:48:38 UTC (60 KB)
来源：arXiv：cs.LG · arxiv.org