EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Personalized anomaly detection using deep active learning

Alireza Vafaei Sadr, Bruce A Bassett, Emmanuel Sekyi

RAS Techniques and Instruments · 17.01.2023 · 2 (1) · 586-598

DOI: 10.1093/rasti/rzad032 ↗

Kurz gefasst

Fachartikel in der Zeitschrift RAS Techniques and Instruments, online erschienen am 06.09.2023. Verfasst von Alireza Vafaei Sadr und 2 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 22 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: University of Geneva.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Anomaly detection algorithms are typically applied to static, unchanging, data features hand-crafted by the user. But how does a user systematically craft good features for anomalies that have never been seen? Here we couple deep learning with active learning – in which an Oracle iteratively labels small amounts of data selected algorithmically over a series of rounds – to automatically and dynamically improve the data features for efficient outlier detection. This approach, Anomaly Hunt (ahunt), shows excellent performance on MNIST, CIFAR10, and Galaxy-DECaLS data, significantly outperforming both standard anomaly detection and active learning algorithms with static feature spaces. Beyond improved performance, ahunt also allows the number of anomaly classes to grow organically in response to the Oracle’s evaluations. Extensive ablation studies explore the impact of Oracle question selection strategy and loss function on performance. We illustrate how the dynamic anomaly class taxonomy represents another step towards fully personalized rankings of different anomaly classes that reflect a user’s interests, allowing the algorithm to learn to ignore statistically significant but uninteresting outliers (e.g. noise). This should prove useful in the era of massive astronomical data sets serving diverse sets of users who can only review a tiny subset of the incoming data.

Förderung

  • University of Geneva

Zitieren

APA

Sadr, A. V., Bassett, B. A., & Sekyi, E. (2023). Personalized anomaly detection using deep active learning. RAS Techniques and Instruments, 2(1), 586-598. https://doi.org/10.1093/rasti/rzad032

Vancouver

Sadr AV, Bassett BA, Sekyi E. Personalized anomaly detection using deep active learning. RAS Techniques and Instruments. 2023;2(1):586-598. doi:10.1093/rasti/rzad032

BibTeX
@article{Sadr2023,
  author = {Sadr, Alireza Vafaei and Bassett, Bruce A and Sekyi, Emmanuel},
  title = {Personalized anomaly detection using deep active learning},
  journal = {RAS Techniques and Instruments},
  year = {2023},
  volume = {2},
  number = {1},
  pages = {586--598},
  doi = {10.1093/rasti/rzad032}
}
Kostenlos starten →