Explainable Anomaly Detection on High-Dimensional Time Series Data - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Explainable Anomaly Detection on High-Dimensional Time Series Data

Bijan Rad
  • Fonction : Auteur
  • PersonId : 1122857
Fei Song
  • Fonction : Auteur
  • PersonId : 1122858
Vincent Jacob
Yanlei Diao
  • Fonction : Auteur
  • PersonId : 1052430

Résumé

As enterprise information systems are collecting event streams from various sources, the ability of a system to automatically detect anomalous events and further provide human-readable explanations is of paramount importance. In this paper, we present an approach to integrated anomaly detection (AD) and explanation discovery (ED), which is able to leverage state-of-the-art Deep Learning (DL) techniques for anomaly detection, while being able to recover human-readable explanations for detected anomalies. At the core of the framework is a new human-interpretable dimensionality reduction (HIDR) method that not only reduces the dimensionality of the data, but also maintains a meaningful mapping from the original features to the transformed low-dimensional features. Such transformed features can be fed into any DL technique designed for anomaly detection, and the feature mapping will be used to recover human-readable explanations through a suite of new feature selection and explanation discovery methods. Evaluation using a recent explainable anomaly detection benchmark demonstrates the efficiency and effectiveness of HIDR for AD, and the result that while all three recent ED techniques failed to generate quality explanations on high-dimensional data, our HIDR-based ED framework can enable them to generate explanations with dramatic improvements in the quality of explanations and computational efficiency.
Fichier principal
Vignette du fichier
DEBS2021_HIDR.pdf (723.91 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03522878 , version 1 (12-01-2022)

Identifiants

Citer

Bijan Rad, Fei Song, Vincent Jacob, Yanlei Diao. Explainable Anomaly Detection on High-Dimensional Time Series Data. The 15th ACM International Conference on Distributed and Event-based Systems (DEBS ’21), Jun 2021, virtual event, Italy. ⟨10.1145/3465480.3468292⟩. ⟨hal-03522878⟩
121 Consultations
693 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More