Digital twin driven federated continual learning for predictive modeling of zero day attack evolution in IoT networks

Abstract This rapid evolution of cyberattacks in distributed IoT environment becomes challenging for intrusion detection system which relies on historical attack knowledge, static model, and centralized collected security data. They offers little ability for privacy-preserving collaboration, long-term adaptation and anticipation of unknown new attack behaviors. To address these limitations, a Digital Twin-Driven Federated Continual Learning framework is proposed for predictive zero-day attack evolution modelling. The framework combines digital twin-based attack mutation, privacy-preserving federated learning, continual adaptation, and temporal threat-evolution prediction to model changing adversarial behaviour across distributed IoT environments. Under the revised temporal and attack-aware evaluation protocol, the framework achieved 97.68% detection accuracy, 97.07% F1-score, 92.47% zero-day detection accuracy, and 92.18% attack-evolution prediction accuracy. Performance recovered from 88.92% during concept drift to 94.86% after continual adaptation, while communication analysis indicated a 17.62% reduction compared with reference federated configurations. Future work will consider light-weight edge deployment and better privacy measures. Foundation-model assisted threat reasoning will also be explored together with real world IoT verification.

Authors

Institutions

Publication Details

Journal
Discover Applied Sciences
Published
2026-11-17
DOI
https://doi.org/10.1007/s42452-026-09850-0
Primary Topic
Network Security and Intrusion Detection
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Digital twin driven federated continual learning for predictive modeling of zero day attack evolution in IoT networks

N. Pavithra, M. B. Bharath, B. A. Mala
3 links
Discover Applied Sciences
Network Security and Intrusion Detection
article

Digital twin driven federated continual learning for predictive modeling of zero day attack evolution in IoT networks

N. Pavithra, M. B. Bharath, B. A. Mala
article en

Abstract

Abstract This rapid evolution of cyberattacks in distributed IoT environment becomes challenging for intrusion detection system which relies on historical attack knowledge, static model, and centralized collected security data. They offers little ability for privacy-preserving collaboration, long-term adaptation and anticipation of unknown new attack behaviors. To address these limitations, a Digital Twin-Driven Federated Continual Learning framework is proposed for predictive zero-day attack evolution modelling. The framework combines digital twin-based attack mutation, privacy-preserving federated learning, continual adaptation, and temporal threat-evolution prediction to model changing adversarial behaviour across distributed IoT environments. Under the revised temporal and attack-aware evaluation protocol, the framework achieved 97.68% detection accuracy, 97.07% F1-score, 92.47% zero-day detection accuracy, and 92.18% attack-evolution prediction accuracy. Performance recovered from 88.92% during concept drift to 94.86% after continual adaptation, while communication analysis indicated a 17.62% reduction compared with reference federated configurations. Future work will consider light-weight edge deployment and better privacy measures. Foundation-model assisted threat reasoning will also be explored together with real world IoT verification.

Discover Applied Sciences
Manipal Academy of Higher Education (IN), Dayananda Sagar University (IN)
Network Security and Intrusion Detection
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon