跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.21562 · v1

Severe but Harmless: Stress-Testing Shift Severity for Frozen Time-Series Forecasters

严重但无害:对冻结时间序列预测器进行偏移严重性压力测试

首次提交 2026-11-17 02:40
版本更新 2026-11-17 02:40
雷达观测 2026-11-17 16:10
arXiv ↗PDF ↗

摘要

Zero-shot time-series foundation models (TSFMs) have become attractive alternatives to task-specific models because they handle diverse downstream tasks, including anomaly detection and forecasting, without per-series training. However, time-series data undergo distribution shifts, and the dissimilarity between two regimes is commonly scored with statistical measures such as Wasserstein-1, where a higher score is read as a more severe shift. In systematic experiments using three established severity measures on nine regime shifts from three standard traffic datasets, we show that these measures do not reliably predict degradation for instance-normalized TSFMs. The regimes rated most severe are not the regimes that degrade the models most, and degradation reaches 29% in the lowest-scoring regime. In most cases, the audit shows no positive relationship between measured severity and model degradation. We provide an equivariance-based stress-testing analysis that explains why severity readings can become decoupled from model degradation.

作者备注:10 pages, 6 figures, 5 tables. Accepted at IEEE BigData 2026

AI 论文解读

基于摘要 · 非全文审阅

作者通过系统实验发现,常用的偏移严重性统计度量不能可靠预测实例归一化时间序列基础模型的性能退化,并用等变性压力测试分析解释这种解耦。

正在读取账号与研究记录…

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-17 02:40 · 最新观测版本

Severe but Harmless: Stress-Testing Shift Severity for Frozen Time-Series Forecasters

Zero-shot time-series foundation models (TSFMs) have become attractive alternatives to task-specific models because they handle diverse downstream tasks, including anomaly detection and forecasting, without per-series training. However, time-series data undergo distribution shifts, and the dissimilarity between two regimes is commonly scored with statistical measures such as Wasserstein-1, where a higher score is read as a more severe shift. In systematic experiments using three established severity measures on nine regime shifts from three standard traffic datasets, we show that these measures do not reliably predict degradation for instance-normalized TSFMs. The regimes rated most severe are not the regimes that degrade the models most, and degradation reaches 29% in the lowest-scoring regime. In most cases, the audit shows no positive relationship between measured severity and model degradation. We provide an equivariance-based stress-testing analysis that explains why severity readings can become decoupled from model degradation.

查看此版本解读 →查看此版本原文 ↗