发表机构
Luddy School of Informatics, Computing, and Engineering, Indiana University Bloomington; University of Illinois Springfield; College of Science and Engineering, Hamad Bin Khalifa University（印第安纳大学伯明顿分校勒迪信息学、计算与工程学院; 伊利诺伊大学斯普林菲尔德分校; 哈马德·本·哈利法大学科学与工程学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究揭示时间序列预测中检索插件的收益源于窗口长度与季节周期的关系，而非插件机制，并提出两个统计量在部署前预测收益。
AI 中文摘要
检索插件为深度预测器提供了其回看窗口无法携带的信息。已发表的评估报告显示了一致的改进，且每项改进都归功于其自身的机制。我们表明，这种益处实际上归属于操作点：窗口长度 $S$ 与主导季节周期 $L$ 之间的关系，而标准协议从未改变这一轴。按该关系对评估进行分层，揭示了这一机制。当 $S=12$ 时，一个简单的控制方法——重复最后一个观测周期——在七个基准中的四个上，总体上比六个标准骨干网络在 MSE 上低 $8\%$ 到 $44\%$。它在 ETTm1 上优于我们运行的最强插件，并在 ECL 上与其持平。在三个训练集频谱缺乏集中且共享周期的数据集上，它的表现最多差 $25\%$。对预测范围、周期和窗口的受控合成扫描显示，收益边界与周期（相关性 $+0.71$）而非预测范围（$-0.23$）相关。一个没有相位可恢复的配对控制几乎消除了这种效应，这与相位饥饿一致。零样本预训练无法避免这一点：一个基础模型在周期性基准上比训练过的骨干网络落后 $22\%$ 到 $50\%$。在我们的工具中，精确查找与图扩散相匹配：收益在于查阅记录，而非其上的机制。两个可解释的统计量——趋势检验和过时率——在留一数据集评估下，以 $0.76$ 的准确率预测每个单元的收益符号，相比 $0.69$ 的多数规则具有显著优势，而一个 22 特征堆栈仅达到 $0.57$。我们不提出新的插件。贡献在于机制图、揭示该图的协议，以及在部署前筛选它的两个统计量。代码：此 https URL。
英文摘要
Retrieval plug-ins supply a deep forecaster with information its lookback window cannot carry. Published evaluations report consistent gains, and each credits its own mechanism. We show that the benefit belongs instead to the operating point: the relation between window length $S$ and dominant seasonal period $L$, an axis the standard protocol never varies. Stratifying the evaluation by that relation exposes the regime. At $S{=}12$, a simple control that repeats the last observed period beats the six standard backbones, in aggregate, on four of seven benchmarks by $8\%$ to $44\%$ of MSE. It beats the strongest plug-in we run on ETTm1 and matches it on ECL. It is worse by up to $25\%$ on the three datasets whose training-split spectra lack a concentrated, shared period. A controlled synthetic sweep of horizon, period, and window shows the benefit boundary tracks the period (correlation $+0.71$), not the horizon ($-0.23$). A paired control with no phase to recover nearly erases the effect, consistent with phase starvation. Zero-shot pretraining does not escape it: a foundation model trails trained backbones by $22\%$ to $50\%$ on the periodic benchmarks. Within our instrument, exact lookup matches graph diffusion: the payoff is consulting the record, not the machinery on top. Two interpretable statistics, a trend test and a staleness rate, predict the sign of the per-cell benefit at $0.76$ accuracy under leave-one-dataset-out evaluation, a suggestive margin over the $0.69$ majority rule, where a 22-feature stack manages $0.57$. We propose no new plug-in. The contribution is the regime map, the protocol that reveals it, and two statistics that screen it before deployment. Code: https://github.com/KurbanIntelligenceLab/retrieval-regime.
