Next-Generation Network
Led by Dr Muhammad Nur Firdaus Bin Sahran
SIG Overview
The Next-Generation Network SIG pushes the boundaries of how modern networks are classified, secured, and made equitable. Led by Dr Muhammad Nur Firdaus Bin Sahran, the group tackles the challenge of increasingly complex and encrypted network traffic by harnessing generative AI and adversarial machine learning to build classifiers that stay ahead of emerging threats and data imbalances.
Beyond pure network security, the SIG takes a distinctly human-centred perspective. Research extends to securing health IoT devices for vulnerable communities, and to bridging the digital divide by empowering indigenous communities in Malaysia with the cyber-safety knowledge needed to participate safely in the digital economy. This breadth reflects a conviction that next-generation networks must be both technically resilient and socially inclusive.
Key Research Areas
-
1
Generative AI for Network Traffic Classification
Applying Generative Adversarial Networks (GANs) to synthesise minority-class traffic samples, overcoming the severe class imbalance that degrades classifiers in real-world network datasets. The goal is more robust, production-ready models for traffic identification and anomaly detection.
-
2
Adversarial Machine Learning in Network Security
Investigating how adversarial perturbations and poisoning attacks undermine AI-based security systems, with a focus on generative user-profiling models. Research develops hardened architectures and defence mechanics that remain reliable under deliberate adversarial manipulation.
-
3
Agentic AI for Autonomous Threat Intelligence
Designing protocol-driven agentic AI systems capable of autonomously collecting, correlating, and responding to cyber threats using the Model Context Protocol (MCP). This removes human bottlenecks from the threat-intelligence cycle and enables continuous, real-time defence at scale.
-
4
Digital Inclusion and Indigenous Community Cyber Safety
Developing culturally contextualised cybersecurity education modules for Orang Asal communities in Sabah, Malaysia. Research addresses the unique digital literacy gaps and threat landscapes faced by indigenous populations engaging with e-government, mobile banking, and social media for the first time.
-
5
Health IoT and Wearable Security
Analysing security and privacy risks in smart wearable devices used for remote health monitoring, particularly for stroke rehabilitation. Research examines data integrity, secure transmission protocols, and patient-data governance in resource-constrained IoT environments.
Research Projects
-
1
Generative Adversarial Networks for Mitigating Class Imbalance in Network Traffic Classification
A funded research project (2024-2026) developing GAN-based oversampling techniques tailored for imbalanced network traffic datasets. The work benchmarks novel synthetic data generation strategies against standard resampling baselines on real-world encrypted-traffic corpora.
-
2
Adversarial Machine Learning and Security Mechanics for Generative User Profiling Systems
A nationally-funded project investigating attack vectors against AI-driven user-profiling systems -- particularly those built on large generative models. Outputs include a threat taxonomy, adversarial benchmarks, and a countermeasure framework for deployment-hardened profiling pipelines.
-
3
Protocol-Driven Agentic AI for Autonomous Threat Analysis and Response
A nationally-funded collaboration designing an end-to-end agentic AI architecture that leverages MCP to orchestrate threat-hunting agents across heterogeneous security toolchains. The system is evaluated against real incident-response scenarios to measure detection latency and false-positive reduction.
-
4
Shielding Our Heritage: Digital Safety for Orang Asal Sabah
A community-engaged project (2026) co-designing cyber-safety modules with indigenous communities in Sabah. Modules cover phishing awareness, social-media privacy, and safe use of digital government services -- delivered in formats accessible to low-literacy and multilingual audiences.
-
5
Smart Wearable Technology for Health Monitoring Among Stroke Survivors
A cross-disciplinary project (2024-2026) examining how wearable health devices can securely support stroke rehabilitation at home. Research evaluates the security posture of commercial wearables, proposes lightweight encryption schemes for biometric telemetry, and assesses patient trust and data-governance requirements.
Interested in collaborating?
Reach out to discuss postgraduate supervision, joint research, or industry partnerships in next-generation network security and AI-driven threat intelligence.
Get in Touch