PaperarXiv:2611.14425
Quantifying the Sampling Frequency Invariance of Time-Series Foundation Models for Sensor Data Applications
arXiv preprint (cs.LG)
TL;DR
Bosch / Georgia Tech stress test of MOMENT, Chronos-2, Moirai-MoE and TimesFM under sampling-frequency changes, with three new metrics (FSI, FDI, CDI). Invariance depends mainly on the margin between a signal's native rate and the bandwidth the task needs, not on which TSFM you pick.
Abstract
Time-Series Foundation Models (TSFMs) are increasingly used in sensor applications, where sampling frequency (SF) varies across sensors, devices, and operating modes (e.g., power saving). Ideally, TSFMs must remain effective during such SF changes, provided the SF retains task-critical bandwidth. In this paper, we stress-test 4 prominent TSFMs: MOMENT, Chronos-2, Moirai-MoE, and TimesFM, to quantify their SF-invariance and investigate whether their behavior tracks the resulting loss of information bandwidth rather than the SF value itself. Our stress-test probes 3 dimensions: (i) the shift in embeddings, (ii) the degradation in forecasting performance, and (iii) the degradation in downstream performance, quantified by three novel metrics: FSI, FDI, and CDI, respectively. To our knowledge, this is the first study to place SF-invariance at the core of TSFM evaluations. Our results characterize: (i) how SF-invariance depends on the margin between signal's native-SF and the bandwidth its task requires, (ii) how much of that invariance the choice of TSFM explains, and (iii) how much of TSFM's performance degradation under a lower SF reflects information loss rather than a limitation of their abilities. These findings reframe SF-invariance as a measurable, signal-aware property of the (signal X TSFM) pair, and guide practitioners on when TSFMs display SF-invariance.
Verbatim from arXiv (opens in a new tab) · arXiv metadata (abstracts are released under CC0)
Results on ChatPPG leaderboards
All leaderboardsNo results from this paper on ChatPPG leaderboards yet. Browse the leaderboards.
Tasks & signals
Tasks
Signals
Keywords
Datasets used
How it works
Cite this paper
Dhekane, S. G., Kumar, A., Woertwein, T., Lin, Y.-B., Ploetz, T., & Ghaffarzadegan, S. (2026). Quantifying the Sampling Frequency Invariance of Time-Series Foundation Models for Sensor Data Applications. arXiv preprint. https://arxiv.org/abs/2611.14425
Dhekane, Sourish Gunesh, Abinaya Kumar, Torsten Woertwein, Yan-Bo Lin, Thomas Ploetz, and Shabnam Ghaffarzadegan. 2026. "Quantifying the Sampling Frequency Invariance of Time-Series Foundation Models for Sensor Data Applications." arXiv preprint. https://arxiv.org/abs/2611.14425.
Dhekane, S. G., Kumar, A., Woertwein, T., Lin, Y.-B., Ploetz, T. and Ghaffarzadegan, S. (2026) 'Quantifying the Sampling Frequency Invariance of Time-Series Foundation Models for Sensor Data Applications', arXiv preprint. Available at: https://arxiv.org/abs/2611.14425
Dhekane, Sourish Gunesh, et al. "Quantifying the Sampling Frequency Invariance of Time-Series Foundation Models for Sensor Data Applications." arXiv preprint, 2026. https://arxiv.org/abs/2611.14425.
S. G. Dhekane, A. Kumar, T. Woertwein, Y.-B. Lin, T. Ploetz, and S. Ghaffarzadegan, "Quantifying the Sampling Frequency Invariance of Time-Series Foundation Models for Sensor Data Applications," arXiv preprint, 2026, [Online]. Available: https://arxiv.org/abs/2611.14425.
@misc{dhekane2026quantifying,
title = {{Quantifying the Sampling Frequency Invariance of Time-Series Foundation Models for Sensor Data Applications}},
author = {Dhekane, Sourish Gunesh and Kumar, Abinaya and Woertwein, Torsten and Lin, Yan-Bo and Ploetz, Thomas and Ghaffarzadegan, Shabnam},
year = {2026},
howpublished = {arXiv preprint},
eprint = {2611.14425},
archivePrefix = {arXiv}
}Added to the ChatPPG research hub