Research

My research develops trustworthy artificial intelligence for the autonomous management of 5G/6G networks.

Zero-touch network and service management for 5G, beyond-5G and 6G
Network slicing: admission control, SLA assurance, resource dimensioning, O-RAN
Explainable, federated, split and quantum-assisted learning for wireless networks
LLM-based agentic AI for network operations
Network and V2X security: intrusion detection, jamming and poisoning-attack mitigation
MEC and edge–cloud orchestration, IoT/IIoT, vehicular clouds, Industry 4.0

Main contributions

Agentic management

From zero-touch to agentic network management

Agentic-NWDAF, where LLM agents turn operator intents into 3GPP analytics calls (IEEE ICC 2026); an open-source, Release 17-compliant MLOps NWDAF for OpenAirInterface (IEEE ICMLCN 2026); LLM-based explanation of DRL slicing decisions (IEEE NetSoft 2024).

Reliability

Keeping AI reliable over time

The DSA-AE drift detector, with +13.6% accuracy and −94.7% detection latency (IEEE TMLCN 2025); drift adaptation without forgetting (IEEE ICMLCN 2025); uncertainty-aware drift management for DRL (IEEE INFOCOM 2026).

Trust

Trust among network stakeholders

Blockchain and smart-contract architectures for slice SLAs; explainable federated KPI prediction; EUROCOMPLY, agentic auditing against the EU AI Act, GDPR and 3GPP (IEEE Communications Magazine 2026).

Security

Secure, private and lightweight AI

Poisoning-resistant federated learning (IEEE TNSM 2023); early intrusion detection (IEEE TIFS 2024); self-supervised and distilled IDS for 5G-V2X; the DRIVE-B5G testbed.