Graph Equations

Explaining the graph neural networks equations

GCN layer

Semi-Supervised Classification with Graph Convolutional Networks https://arxiv.org/pdf/1609.02907.pdf

\[ h_i^{(l+1)}=\sigma\left(\sum_{j \in \mathcal{N}_i} \frac{1}{c_{i j}} h_j^{(l)} W^{(l)}\right) \]

where 𝐖(ℓ+) denotes a trainable weight matrix of shape [num_output_features, num_input_features] and 𝑐𝑤,𝑣 refers to a fixed normalization coefficient for each edge.

PyG implements this layer via GCNConv