Anomaly Detection in Industrial Control Networks Using Hybrid LDA - Autoencoder Based Models
Authors
Hua Zhang, Shixiang Zhu, Jun Zhao, Minghui Gao, Zheng Shou, Ye Liang
Corresponding Author
Hua Zhang
Available Online November 2016.
- DOI
- 10.2991/aiea-16.2016.10How to use a DOI?
- Keywords
- Anomaly detection; Industrial Control Networks; LDA; Autoencoder.
- Abstract
This paper introduces a hybrid model that combines Latent Dirichlet Allocation (LDA) model with autoencoder to detect anomalies in Industrial Control Networks. The autoencoder provides a low-dimensional embedding for the input data, whose subsequent distribution is captured by the LDA model. The autoencoder thus acts as a trainable feature extractor while the LDA model captures the group structure of the data. This new approach potentially completes the strength of signature-based and anomaly-based methods.
- Copyright
- © 2016, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Hua Zhang AU - Shixiang Zhu AU - Jun Zhao AU - Minghui Gao AU - Zheng Shou AU - Ye Liang PY - 2016/11 DA - 2016/11 TI - Anomaly Detection in Industrial Control Networks Using Hybrid LDA - Autoencoder Based Models BT - Proceedings of the 2016 International Conference on Artificial Intelligence and Engineering Applications PB - Atlantis Press SP - 53 EP - 58 SN - 2352-538X UR - https://doi.org/10.2991/aiea-16.2016.10 DO - 10.2991/aiea-16.2016.10 ID - Zhang2016/11 ER -