arXiv cs.AI· Laurent Mombaerts, Jonathan Taws, Jacopo Gargano, Dinesh Gaddam, Kevin George, Srijan Tiwari, Shreyas Rajesh·· 12 小时前AI 评分31
无需训练的多模态预测持续学习：用提示词进化替代重训练
Training-Free Continual Learning for Multimodal Forecasting via Prompt Evolution
AI 导读
研究者提出用一段由结果驱动反思进化出的文本引导块，替代重训练来让冻结的 LLM 预测器持续适应新数据分布。在严格防数据泄漏的多模态基准上，Llama-3.3-70B 携带 900 词进化引导取得 GM MASE 0.776，优于 TabPFN-TS-3 的 0.898，且可无损迁移到更新模型，更新成本为美元级而非 GPU-days。
正文 · 原文
Abstract:Adapting a forecasting system to a changing world classically means retraining. We ask whether the task of continually learning regimes for time series forecasting can be represented by a piece of text: a guidance block, evolved by outcome-driven reflection and used in a frozen LLM forecaster's prompt. On a refreshed multimodal benchmark (weekly search indices, daily currencies, daily commodities, each window paired with web-retrieved news), we run GEPA-style prompt evolution under strict contamination control to prevent the data leakage common in timeseries benchmarks: the forecaster's knowledge cutoff predates every evaluated window. A 900-word evolved guidance carried by Llama-3.3-70B beats TimeSeries Foundation Models (TSFMs) on held-out data (GM MASE 0.776 vs. 0.898 for TabPFN-TS-3), transfers to more recent models without measurable loss, and costs dollars rather than GPU-days to update. We characterize the loop: out-of-loop gains concentrate in the first reflection, in-loop improvements overstate out-of-loop gains 2-3x unless an untouched monitor arbitrates, and error-only objectives silently suppress the text modality. We provide two extensions. First, compiling the guidance into executable code with a TSFM as a callable tool recovers most of the gain deterministically. Second, ablations show the advantage widens at short horizons and that removing the news text does not hurt the guided forecaster, consistent with the text's value being distilled into policy during learning. We argue that continual prompt evolution is a practical, training-free alternative to retraining for multimodal forecasting.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.27468 [cs.LG] |
| (or arXiv:2610.27468v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.27468 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Dinesh Yadav Gaddam [view email]
[v1]
Tue, 20 Oct 2026 01:38:24 UTC (59 KB)
来源：arXiv cs.AI · arxiv.org