Number of the records: 1
Density-Approximating Neural Network Models for Anomaly Detection
- 1.
SYSNO ASEP 0507118 Document Type C - Proceedings Paper (int. conf.) R&D Document Type Conference Paper Title Density-Approximating Neural Network Models for Anomaly Detection Author(s) Flusser, M. (CZ)
Pevný, T. (CZ)
Somol, Petr (UTIA-B) RIDNumber of authors 3 Source Title ACM SIGKDD 2018 Workshop. - New York : ACM, 2018 - ISBN 978-1-4503-5552-0 Pages s. 1-8 Number of pages 8 s. Publication form Online - E Action ACM SIGKDD 2018 Workshop Event date 20.08.2018 VEvent location London Country GB - United Kingdom Event type WRD Language eng - English Country US - United States Keywords neural network ; anomaly detection Subject RIV BC - Control Systems Theory OECD category Robotics and automatic control Institutional support UTIA-B - RVO:67985556 Annotation We propose an alternative use of neural models in anomaly detection. Traditionally, in anomaly detection context the common use of neural models is in form of auto-encoders. Through the use of auto-encoders the true anomality is proxied by reconstruction error. Auto-encoders often perform well but do not guarantee to perform as expected in all cases. A popular more direct way of modeling anomality distribution is through k-Nearest Neighbor models. Although kNN can perform better than auto-encoders in some cases, their applicability can be seriously impaired by their space and time complexity especially with high-dimensional large-scale data. The alternative we propose is to model the distribution imposed by kNN using neural networks. We show that such neural models are capable of achieving comparable accuracy to kNN while reducing computational complexity by orders of magnitude. The de-noising e ect of a neural model with limited number of neurons and layers is shown to lead to accuracy improvements in some cases. We evaluate the proposed idea against standard kNN and auto-encoders on a large set of benchmark data and show that in majority of cases it is possible to improve on accuracy or computational cost. Workplace Institute of Information Theory and Automation Contact Markéta Votavová, votavova@utia.cas.cz, Tel.: 266 052 201. Year of Publishing 2020
Number of the records: 1