torch_geometric_signed_directed.nn.directed.DIMPA

Classes

DIMPA

The directed mixed-path aggregation model from the

Module Contents

class DIMPA(hop: int, fill_value: float = 0.5)

Bases: torch.nn.Module

The directed mixed-path aggregation model from the DIGRAC: Digraph Clustering Based on Flow Imbalance paper.

Parameters:

  • hop (int) – Number of hops to consider.

  • fill_value (float, optional) – The layer computes \(\mathbf{\hat{A}}\) as \(\mathbf{A} + fill_value*\mathbf{I}\). (default: 0.5)

conv_layer
forward(x_s: torch.FloatTensor, x_t: torch.FloatTensor, edge_index: torch.FloatTensor, edge_weight: torch.FloatTensor) → torch.FloatTensor

Making a forward pass of DIMPA.

Arg types:

  • x_s (PyTorch FloatTensor) - Souce hidden representations.

  • x_t (PyTorch FloatTensor) - Target hidden representations.

  • edge_index (PyTorch FloatTensor) - Edge indices.

  • edge_weight (PyTorch FloatTensor) - Edge weights.

Return types:

  • feat (PyTorch FloatTensor) - Embedding matrix, with shape (num_nodes, 2*input_dim).