发表机构
Supcon Technology; College of Artificial Intelligence, Zhejiang University（中控技术; 浙江大学人工智能学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对时间序列模型，分析了延迟可恢复性、报告与功能使用的问题，发现多数模型存在延迟报告正确但实际未使用对应历史的情况，提出的路由方法可实现预测与报告的对齐。
AI 中文摘要
预测准确率无法告知我们是哪些过去的输入产生了预测结果。对于具有已知延迟结构的时间序列模型，我们区分三个问题：能否从观测数据中恢复真实延迟，模型是否报告了该延迟，以及预测是否实际使用了相同的历史信息。我们首先推导了输入条件下的可恢复性度量，该度量将内在歧义与模型误差区分开。随后我们证明，即使预测器仍使用错误的滞后值，延迟报告也能变得任意可靠，同时预测风险接近最优（oracle）。这种失效也出现在点延迟任务的有限样本中：在延迟报告正确且归一化超额风险处于最优值10%以内的预测中，根据我们的匹配掩码测试，N-HiTS案例中有55.4%、TCN案例中有92.7%的预测，其报告的历史信息未被功能使用。最后，我们表明通过报告的历史信息路由预测可消除报告外的旁路路径；硬独热控制可实现精确的固定报告对齐。主要结论很简单：即使延迟报告正确，良好的预测也不意味着模型使用了正确的历史信息。
英文摘要
Forecast accuracy is often used as a proxy for temporal structure discovery, but predictive performance and structural identifiability are not equivalent. Different temporal mechanisms can achieve similar forecast errors, while small forecast differences may still contain sufficient information for recovery. In this work, we study when forecast-only structural selection can be trusted. We show that a vanishing forecast margin does not necessarily imply structural ambiguity, and establish a stability perspective that evaluates structural separation relative to uncertainty in the selection objective. This perspective provides both a sufficient condition for reliable selection and a continuous measure of selection difficulty. Experiments across controlled and end-to-end settings demonstrate that forecast margin alone is insufficient, while the proposed stability measure better characterizes when forecast-based structural selection succeeds or fails. Our results suggest that predictive accuracy should be treated as evidence for structure discovery only when its separation is sufficiently robust.
