Frequency Matters: When Time Series Foundation Models Fail Under Spectral Shift
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908637844209664 |
|---|---|
| author | Wang, Tianze Ennadir, Sofiane Pertoft, John Gandler, Gabriela Zarzar Cao, Lele Senane, Zineb Katsarou, Styliani Asadi, Sahar Karlsson, Axel Smirnov, Oleg |
| author_facet | Wang, Tianze Ennadir, Sofiane Pertoft, John Gandler, Gabriela Zarzar Cao, Lele Senane, Zineb Katsarou, Styliani Asadi, Sahar Karlsson, Axel Smirnov, Oleg |
| contents | Time series foundation models (TSFMs) have shown strong results on public benchmarks, prompting comparisons to a "BERT moment" for time series. Their effectiveness in industrial settings, however, remains uncertain. We examine why TSFMs often struggle to generalize and highlight spectral shift (a mismatch between the dominant frequency components in downstream tasks and those represented during pretraining) as a key factor. We present evidence from an industrial-scale player engagement prediction task in mobile gaming, where TSFMs underperform domain-adapted baselines. To isolate the mechanism, we design controlled synthetic experiments contrasting signals with seen versus unseen frequency bands, observing systematic degradation under spectral mismatch. These findings position frequency awareness as critical for robust TSFM deployment and motivate new pretraining and evaluation protocols that explicitly account for spectral diversity. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_05619 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Frequency Matters: When Time Series Foundation Models Fail Under Spectral Shift Wang, Tianze Ennadir, Sofiane Pertoft, John Gandler, Gabriela Zarzar Cao, Lele Senane, Zineb Katsarou, Styliani Asadi, Sahar Karlsson, Axel Smirnov, Oleg Machine Learning Artificial Intelligence Time series foundation models (TSFMs) have shown strong results on public benchmarks, prompting comparisons to a "BERT moment" for time series. Their effectiveness in industrial settings, however, remains uncertain. We examine why TSFMs often struggle to generalize and highlight spectral shift (a mismatch between the dominant frequency components in downstream tasks and those represented during pretraining) as a key factor. We present evidence from an industrial-scale player engagement prediction task in mobile gaming, where TSFMs underperform domain-adapted baselines. To isolate the mechanism, we design controlled synthetic experiments contrasting signals with seen versus unseen frequency bands, observing systematic degradation under spectral mismatch. These findings position frequency awareness as critical for robust TSFM deployment and motivate new pretraining and evaluation protocols that explicitly account for spectral diversity. |
| title | Frequency Matters: When Time Series Foundation Models Fail Under Spectral Shift |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2511.05619 |