Temporal-Conditioned Normalizing Flows for Multivariate Time Series Anomaly Detection
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910047538249728 |
|---|---|
| author | Baumgartner, David Langseth, Helge Engø-Monsen, Kenth Ramampiaro, Heri |
| author_facet | Baumgartner, David Langseth, Helge Engø-Monsen, Kenth Ramampiaro, Heri |
| contents | This paper introduces temporal-conditioned normalizing flows (tcNF), a novel framework that addresses anomaly detection in time series data with accurate modeling of temporal dependencies and uncertainty. By conditioning normalizing flows on previous observations, tcNF effectively captures complex temporal dynamics and generates accurate probability distributions of expected behavior. This autoregressive approach enables robust anomaly detection by identifying low-probability events within the learned distribution. We evaluate tcNF on diverse datasets, demonstrating good accuracy and robustness compared to existing methods. A comprehensive analysis of strengths and limitations and open-source code is provided to facilitate reproducibility and future research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_09490 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Temporal-Conditioned Normalizing Flows for Multivariate Time Series Anomaly Detection Baumgartner, David Langseth, Helge Engø-Monsen, Kenth Ramampiaro, Heri Machine Learning Artificial Intelligence This paper introduces temporal-conditioned normalizing flows (tcNF), a novel framework that addresses anomaly detection in time series data with accurate modeling of temporal dependencies and uncertainty. By conditioning normalizing flows on previous observations, tcNF effectively captures complex temporal dynamics and generates accurate probability distributions of expected behavior. This autoregressive approach enables robust anomaly detection by identifying low-probability events within the learned distribution. We evaluate tcNF on diverse datasets, demonstrating good accuracy and robustness compared to existing methods. A comprehensive analysis of strengths and limitations and open-source code is provided to facilitate reproducibility and future research. |
| title | Temporal-Conditioned Normalizing Flows for Multivariate Time Series Anomaly Detection |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2603.09490 |