Robust and Explainable Detector of Time Series Anomaly via Augmenting Multiclass Pseudo-Anomalies Conference: Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2; August 3–7, 2025; Toronto, ON, Canada Proceedings of the 31st ACM SIGKDD Conference on Knowledge Dis
Kohei Obata, Yasuko Matsubara, Yasushi Sakurai · 2025
Read Robust and Explainable Detector of Time Series Anomaly via Augmenting Multiclass Pseudo-Anomalies Conference: Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2; August 3–7, 2025; Toronto, ON, Canada Proceedings of the 31st ACM SIGKDD Conference on Knowledge Dis by Kohei Obata, Yasuko Matsubara, Yasushi Sakurai (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-20765.txt.
Description
Kohei Obata, Yasuko Matsubara, and Yasushi Sakurai hunt anomalies in time series without trusting the training set. Their 2025 paper notes that real training data can hide anomalies, which teach a model the wrong normal. Some rivals fake anomalies by augmentation, yet their fakes span too few kinds, waste near-normal fakes, and trust labels too hard. Their system, RedLamp, fakes many classes of pseudo-anomalies with diverse twists.
A multiclass sort with soft labels learns the boundary between real and fake classes. Soft labels curb overconfidence and add grit against stray real anomalies. The learned space doubles as an explanation, since fake kinds stay apart. Wide tests back the gains.
Catalog record
| CREATOR | Kohei Obata, Yasuko Matsubara, Yasushi Sakurai |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 58 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-20765