SCIENCE@home

Score and You Shall Find: A Novel Regularization Technique for Cyberattack Detection in Industrial Control Systems

Tatsumi Oba,Tadahiro Taniguchi,Naoto Yanai

Published 2023 in RICSS@CCS

ABSTRACT

In recent years, countermeasures against cyberattacks on industrial control systems (ICSs) has included the use of machine learning models for their detection. However, the efficacy of these existing detection methods is limited due to a large number of false positives. In this paper, we propose a novel regularization technique, named similar device regularization, to reduce false positives of cyberattack detection in ICSs. The proposed technique penalizes the separation of feature vectors for similar devices and is applicable to any machine learning model with link prediction tasks for cyberattack detection in ICSs. Furthermore, we present a detection method, ConvSDR, as an application of the proposed technique. Extensive experiments with ConvSDR demonstrate that it outperforms the existing methods by virtue of the similar device regularization. The similar device regularization can also suppress the overfitting of machine learning models, which is often caused due to an increase in model parameters. Furthermore, we identify that the similar device regularization reduces false positives by over 40%.

PUBLICATION RECORD

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-42 of 42 references · Page 1 of 1

CITED BY