| Title: |
Supervised aggregation of anomaly score functions for active anomaly detection |
| Authors: |
Bleakley, Kevin; Mendil, Mouhcine; Royer, Martin; Auder, Benjamin |
| Contributors: |
Statistique mathématique et apprentissage (CELESTE); Laboratoire de Mathématiques d'Orsay (LMO); Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS)-Centre Inria de l'Université Paris-Saclay; Centre Inria de Saclay; Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Centre Inria de Saclay; Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria); Université Paris-Saclay; Université Paris-Saclay-Centre National de la Recherche Scientifique (CNRS); IRT Saint Exupéry - Institut de Recherche Technologique; Understanding the Shape of Data (DATASHAPE); Centre Inria d'Université Côte d'Azur; Institut National de Recherche en Informatique et en Automatique (Inria)-Institut National de Recherche en Informatique et en Automatique (Inria)-Laboratoire de Mathématiques d'Orsay (LMO); Institut National de Recherche en Informatique et en Automatique (Inria); IRT SystemX; This work was supported by the French government under the "France 2030" program via INRIA as part of the Confiance.ai project. |
| Source: |
ISSN: 2835-8856 ; Transactions on Machine Learning Research Journal ; https://inria.hal.science/hal-05059280 ; Transactions on Machine Learning Research Journal, In press. |
| Publisher Information: |
CCSD; OpenReview.net, 2022 |
| Publication Year: |
2026 |
| Subject Terms: |
[MATH]Mathematics [math] |
| Description: |
International audience ; Detecting rare anomalies in batches of multidimensional data is challenging. We propose a supervised active-learning framework that sends a small number of data points from each batch to an expert for labeling as 'anomaly' or 'nominal', via two mechanisms: (i) points most likely to be an anomaly in the eyes of a supervised classifier trained on previously-labeled data; and (ii) points suggested by an active learner. Instead of, however, training the supervised classifier directly on the current set of labeled raw data points, we treat the scores calculated by an ensemble of M unsupervised anomaly detectors on each data point as if they were the learner's input features. This approach generalizes earlier attempts to linearly aggregate unsupervised anomaly detector scores, and broadens the scope of such methods to ordered data like time series. Results suggest that this method usually outperforms-often significantly-linear strategies. The Python library acanag provides an implementation of the proposed method. |
| Document Type: |
article in journal/newspaper |
| Language: |
English |
| Availability: |
https://inria.hal.science/hal-05059280; https://inria.hal.science/hal-05059280v3/document; https://inria.hal.science/hal-05059280v3/file/AAApreprintV3.pdf; https://inria.hal.science/hal-05059280v3/file/AAAappendixV3.pdf |
| Rights: |
https://creativecommons.org/licenses/by/4.0/ ; info:eu-repo/semantics/OpenAccess |
| Accession Number: |
edsbas.EB2F3BCC |
| Database: |
BASE |