Graph Neural Networks
Graph Neural Networks (GNNs) extend deep learning to graph-structured data by aggregating neighborhood information through message passing mechanisms.
Contents:
- GCN — Graph convolutional networks, spectral methods, message passing framework
- GraphSAGE — Sampling and aggregation, inductive learning
- GAT — Graph attention networks, multi-head attention
- GNN Applications — Molecular design, recommender systems, knowledge graphs, traffic prediction