Deep actor-critic reinforcement learning for anomaly detection

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations

Abstract

Anomaly detection is widely applied in a variety of domains, involving for instance, smart home systems, network traffic monitoring, IoT applications and sensor networks. In this paper, we study deep reinforcement learning based active sequential testing for anomaly detection. We assume that there is an unknown number of abnormal processes at a time and the agent can only check with one sensor in each sampling step. To maximize the confidence level of the decision and minimize the stopping time concurrently, we propose a deep actor-critic reinforcement learning framework that can dynamically select the sensor based on the posterior probabilities. We provide simulation results for both the training phase and testing phase, and compare the proposed framework with the Chernoff test in terms of claim delay and loss.

Original languageEnglish (US)
Title of host publication2019 IEEE Global Communications Conference, GLOBECOM 2019 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728109626
DOIs
StatePublished - Dec 2019
Event2019 IEEE Global Communications Conference, GLOBECOM 2019 - Waikoloa, United States
Duration: Dec 9 2019Dec 13 2019

Publication series

Name2019 IEEE Global Communications Conference, GLOBECOM 2019 - Proceedings

Conference

Conference2019 IEEE Global Communications Conference, GLOBECOM 2019
CountryUnited States
CityWaikoloa
Period12/9/1912/13/19

Keywords

  • Actor-critic framework
  • Anomaly detection
  • Deep reinforcement learning

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Hardware and Architecture
  • Information Systems
  • Signal Processing
  • Information Systems and Management
  • Safety, Risk, Reliability and Quality
  • Media Technology
  • Health Informatics

Fingerprint Dive into the research topics of 'Deep actor-critic reinforcement learning for anomaly detection'. Together they form a unique fingerprint.

  • Cite this

    Zhong, C., Gursoy, M. C., & Velipasalar, S. (2019). Deep actor-critic reinforcement learning for anomaly detection. In 2019 IEEE Global Communications Conference, GLOBECOM 2019 - Proceedings [9013223] (2019 IEEE Global Communications Conference, GLOBECOM 2019 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/GLOBECOM38437.2019.9013223