🔗 LinkedIn пост · Ishfaq Bashir (PhD) · Researcher in Generative AI & Cybersecurity. · 2026-05-27 · 👍 23 · 💬 2 · 🔁 0
Glad to share that our paper titled “LIME: Lightweight Multi-Head Early-Exit Transformer for Network Intrusion Detection” presented at the IEEE ICC 2026. Glasgow, Scotland, UK.
This work focuses on designing a lightweight and efficient Transformer-based Intrusion Detection System (IDS) for real-time and resource-constrained environments such as Edge and IoT networks.
The proposed LIME framework introduces: • Record Projection Encoding (RPE) for compact feature representation • Lightweight Transformer architecture for efficient contextual learning • Multi-Head Early-Exit inference for adaptive low-latency detection
Experimental results demonstrated strong minority attack detection performance with significantly reduced computational overhead across benchmark IDS datasets.
Grateful to my supervisors, collaborators, and everyone who contributed to this research work.
Special thanks to Dr Silvirianti Silvirianti for kindly presenting the paper on my behalf.
ieeeicc2026 cybersecurity intrusiondetection artificialintelligence deeplearning transformer iot edgecomputing networksecurity research icc