Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Activation Anomaly Analysis

: Sperl, Philip; Schulze, Jan-Philipp; Böttinger, Konstantin


Hutter, Frank:
Machine Learning and Knowledge Discovery in Databases. European Conference, ECML PKDD 2020. Proceedings. Pt.II : Ghent, Belgium, September 14-18, 2020
Cham: Springer Nature, 2021 (Lecture Notes in Artificial Intelligence 12458)
ISBN: 978-3-030-67661-2
European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD) <2020, Online>
Bundesministerium für Bildung und Forschung BMBF (Deutschland)
16KIS0933K; IUNO InSec
Fraunhofer AISEC ()
anomaly detection; deep learning; intrusion detection; semi-supervised learning; coverage analysis; data mining; it security

Inspired by recent advances in coverage-guided analysis of neural networks, we propose a novel anomaly detection method. We show that the hidden activation values contain information useful to distinguish between normal and anomalous samples. Our approach combines three neural networks in a purely data-driven end-to-end model. Based on the activation values in the target network, the alarm network decides if the given sample is normal. Thanks to the anomaly network, our method even works in semi-supervised settings. Strong anomaly detection results are achieved on common data sets surpassing current baseline methods. Our semi-supervised anomaly detection method allows to inspect large amounts of data for anomalies across various applications.