NeuroGNN is a state-of-the-art framework for precise seizure detection and classification from EEG data. It employs dynamic Graph Neural Networks (GNNs) to capture intricate spatial, temporal, semantic, and taxonomic correlations between EEG electrode locations and brain regions, resulting in improved accuracy. Presented at PAKDD '24.

Dynamic Graph Neural Networks to Predict Collaborations between Authors