NEXUS-MI

Sep 9, 2026 · 1 min read
projects

NEXUS-MI is a gateway-coordinated federated learning framework for personalized motor-imagery brain-computer interfaces. It supports shared-backbone learning and local classifier-head adaptation while keeping raw EEG and personalized heads local. Its communication policies govern gateway participation, offline buffering, stale updates, and backbone distribution, and are evaluated through public EEG dataset replay.

Read the paper.

Brain-Computer Interfaces Federated Learning Edge Computing Communication-Aware Learning
Daniel Adu Worae
Authors
Daniel Adu Worae (he/him)
Ph.D. Candidate in Computer Science and Engineering
I am a Ph.D. Candidate in Computer Science and Engineering at the University of Notre Dame. My research focuses on secure, interpretable, and privacy-preserving IoT systems, with emphasis on network traffic analysis, anomaly detection, retrieval-augmented generation, and LLM-augmented operational intelligence.