Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
August 13, 2026Discover Internet of ThingsOpen Access

A hybrid CNN-BiLSTM edge-cloud intrusion detection system with online incremental learning and SHAP explainability for smart city IoT

View Full Paper
Ask AI
Bookmark
Share

Discussion

Loading...

Member takes

Implication

Randomized trial demonstrates improved detection accuracy in smart city IoT, suggesting enhanced security analytics.

Key Points

  • The aim is to develop a robust hybrid intrusion detection system for smart city IoT environments that can handle evolving threats and provide explainability.
  • Developed MI-IDS using a hybrid CNN-BiLSTM architecture within an edge-cloud framework.
  • Implemented reservoir-sampling-based incremental learning and SHAP explainability.
  • Conducted evaluations on a benchmark of 73,100 instances and UNSW-NB15 with structured ablation studies.
  • Achieved 96.7% accuracy and 95.8% F1-score on a composite benchmark.
  • Under concept drift simulation, accuracy loss limited to 1.2% for MI-IDS compared to 20.4% for the baseline.
  • SHAP reduced analyst mean time-to-decision by 57.1% in a preliminary study.

Cite This Study

A 2026 study studied this question.

synapsesocial.com/papers/6a7d75d82b0e0cff3f63ec64https://doi.org/10.1007/s43926-026-00459-8
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Enhancing IoT network security with explainable deep learning-based intrusion detection systems2026
  2. 2Scalable and Interpretable Deep Learning‐Based Intrusion Detection Framework for Secure Internet of Things Networks2026 · 3 citations
  3. 3Explainable and Trustworthy Network Intrusion Detection: From Multi-Class SHAP Analysis to Confidence-Aware Deployment2026
  4. 4XAI-Enhanced adversarial resilient deep learning framework for transparent and secure edge deployment in consumer Internet of Things/Industrial Internet of Things environments2026 · 10 citations
  5. 5HybridML CyberShield for explainable proactive intrusion detection in enterprise and IoT networks2026