MARICYBER-6G: AI-Driven Marine Cybersecurity for 6G-Enabled Underwater Wireless Sensor Networks

MARICYBER-6G: AI-Driven Marine Cybersecurity for 6G-Enabled Underwater Wireless Sensor Networks Exhibited at the Cyber Security Conference of the Irish EU Presidency and Cyber Expo (NCSC Conference) , Convention Centre Dublin, 6–8 October 2026 (Cyber Expo). Project: MARI-CYBERWISE Fellowship: Government of Ireland Postdoctoral Fellowship (GOIPD/2025/1090) Affiliation: Department of Electronic and Computer Engineering, University of Limerick, Ireland This repository contains the editable LaTeX source, compiled poster, research figures, references, and supporting assets for the A0 portrait research poster MARICYBER-6G: AI-Driven Marine Cybersecurity for 6G-Enabled Underwater Wireless Sensor Networks. The poster presents an artificial intelligence-driven cybersecurity framework for 6G-enabled Underwater Wireless Sensor Networks (UWSNs), integrating simulation-driven cyber-threat traffic generation, heterogeneous feature construction, deep learning-based intrusion detection, and adaptive mitigation. It examines Normal, distributed denial-of-service (DDoS), and Spoofing traffic in a three-dimensional underwater network comprising 500 nodes within a 1000×1000×1000 m³ volume. The proposed Residual Temporal Attention Network with Bidirectional Long Short-Term Memory (ReTAN-BiLSTM) combines multi-scale one-dimensional convolution, residual feature refinement, bidirectional temporal modelling, temporal attention, and raw-feature fusion to classify the three traffic classes. The poster includes the UWSN topology, feature-correlation structure, network performance under different traffic conditions, model architecture, learning dynamics, class-wise training performance, receiver operating characteristic (ROC) analysis, convergence plots, confusion matrices, and adaptive mitigation results. ReTAN-BiLSTM achieves reported training and test ROC-AUC values of 0.9911 and 0.9868, respectively. Repository Contents Editable A0 portrait LaTeX poster source (main.tex) Compiled poster PDF Research figures and visualisations, including: ReTAN-BiLSTM architecture UWSN topology and feature-correlation figures Network-performance and adaptive-mitigation results Model learning, class-wise performance, and convergence plots ROC curves and confusion matrices University of Limerick, NCSC, Research Ireland, and Irish Research Council logo assets README documentation and compilation instructions