A Tsetlin Machine-driven Intrusion Detection System for Next-Generation IoMT Security
2026-04-03 • Cryptography and Security
Cryptography and SecurityMachine Learning
AI summaryⓘ
The authors designed a new system to spot cyberattacks on connected medical devices that communicate over the Internet. Their system uses a method called the Tsetlin Machine, which is a type of machine learning that creates easy-to-understand rules to recognize attack patterns. Tested on a large diverse dataset, their method was more accurate than traditional models in telling safe activity from attacks and also identifying different attack types. They also made the system explain its decisions by showing which rules and parts of the data were most important. This helps users trust and understand how it works.
Internet of Medical Things (IoMT)Intrusion Detection System (IDS)Tsetlin Machinemachine learningcybersecuritypropositional logicCICIoMT-2024 datasetbinary classificationmulti-class classificationmodel interpretability
Authors
Rahul Jaiswal, Per-Arne Andersen, Linga Reddy Cenkeramaddi, Lei Jiao, Ole-Christoffer Granmo
Abstract
The rapid adoption of the Internet of Medical Things (IoMT) is transforming healthcare by enabling seamless connectivity among medical devices, systems, and services. However, it also introduces serious cybersecurity and patient safety concerns as attackers increasingly exploit new methods and emerging vulnerabilities to infiltrate IoMT networks. This paper proposes a novel Tsetlin Machine (TM)-based Intrusion Detection System (IDS) for detecting a wide range of cyberattacks targeting IoMT networks. The TM is a rule-based and interpretable machine learning (ML) approach that models attack patterns using propositional logic. Extensive experiments conducted on the CICIoMT-2024 dataset, which includes multiple IoMT protocols and cyberattack types, demonstrate that the proposed TM-based IDS outperforms traditional ML classifiers. The proposed model achieves an accuracy of 99.5\% in binary classification and 90.7\% in multi-class classification, surpassing existing state-of-the-art approaches. Moreover, to enhance model trust and interpretability, the proposed TM-based model presents class-wise vote scores and clause activation heatmaps, providing clear insights into the most influential clauses and the dominant class contributing to the final model decision.