galax.nn.zoo.gat

Graph Attention Network

Classes

GAT(features, num_heads, feat_drop, ...)

Apply Graph Attention Network over an input signal. .. math:: h_i^{(l+1)} = sum_{jin mathcal{N}(i)} alpha_{i,j} W^{(l)} h_j^{(l)} where \(\alpha_{ij}\) is the attention score bewteen node \(i\) and node \(j\): .. math:: alpha_{ij}^{l} &= mathrm{softmax_i} (e_{ij}^{l}) e_{ij}^{l} &= mathrm{LeakyReLU}left(vec{a}^T [W h_{i} | W h_{j}]right).