torch_geometric_signed_directed.utils.signed.link_sign_loss

Classes

Sign_Triangle_Loss	An implementation of the Signed Triangle Loss used in
Sign_Direction_Loss	An implementation of the Signed Direction Loss used in
Sign_Product_Entropy_Loss	An implementation of the Signed Entropy Loss used in
Link_Sign_Product_Loss	An implementation of the Product Loss used in
Link_Sign_Entropy_Loss	An implementation of the Entropy Loss used in
Sign_Structure_Loss

Module Contents

class Sign_Triangle_Loss(emb_dim: int, edge_weight: scipy.sparse.csc_matrix)

Bases: torch.nn.Module

An implementation of the Signed Triangle Loss used in

“SDGNN: Learning Node Representation for Signed Directed Networks” paper.

Parameters:

emb_dim (int) – The embedding size.

lin

edge_weight

forward(z: torch.Tensor, pos_edge_index: torch.LongTensor, neg_edge_index: torch.LongTensor) → torch.Tensor

class Sign_Direction_Loss(emb_dim: int)

Bases: torch.nn.Module

An implementation of the Signed Direction Loss used in

“SDGNN: Learning Node Representation for Signed Directed Networks” paper.

Parameters:

emb_dim (int) – The embedding size.

score_function1

score_function2

forward(z: torch.Tensor, pos_edge_index: torch.LongTensor, neg_edge_index: torch.LongTensor) → torch.Tensor

class Sign_Product_Entropy_Loss

Bases: torch.nn.Module

An implementation of the Signed Entropy Loss used in “SDGNN: Learning Node Representation for Signed Directed Networks” paper.

forward(z: torch.Tensor, pos_edge_index: torch.LongTensor, neg_edge_index: torch.LongTensor) → torch.Tensor

class Link_Sign_Product_Loss

Bases: torch.nn.Module

An implementation of the Product Loss used in the “Signed Graph Attention Networks” paper.

forward(z: torch.Tensor, pos_edge_index: torch.LongTensor, neg_edge_index: torch.LongTensor) → torch.Tensor

class Link_Sign_Entropy_Loss(emb_dim: int)

Bases: torch.nn.Module

An implementation of the Entropy Loss used in the “Signed Graph Convolutional Network” paper. and “Learning Signed Network Embedding via Graph Attention” paper

Parameters:

emb_dim (int) – The embedding size.

lin

reset_parameters()

discriminate(z: torch.Tensor, edge_index: torch.LongTensor) → torch.FloatTensor

Given node embeddings z, classifies the link relation between node pairs edge_index to be either positive, negative or non-existent.

Parameters:

• x (Tensor) – The input node features.

• edge_index (LongTensor) – The edge indices.

forward(z: torch.Tensor, pos_edge_index: torch.LongTensor, neg_edge_index: torch.LongTensor) → torch.Tensor

Computes the discriminator loss based on node embeddings z, and positive edges pos_edge_index and negative edges neg_edge_index.

Parameters:

• z (Tensor) – The node embeddings.

• pos_edge_index (LongTensor) – The positive edge indices.

• neg_edge_index (LongTensor) – The negative edge indices.

class Sign_Structure_Loss

Bases: torch.nn.Module

forward(z: torch.Tensor, pos_edge_index: torch.LongTensor, neg_edge_index: torch.LongTensor) → torch.Tensor

pos_embedding_loss(z: torch.Tensor, pos_edge_index: torch.LongTensor) → torch.Tensor

Computes the triplet loss between positive node pairs and sampled non-node pairs.

Parameters:

• z (Tensor) – The node embeddings.

• pos_edge_index (LongTensor) – The positive edge indices.

neg_embedding_loss(z: torch.Tensor, neg_edge_index: torch.LongTensor) → torch.Tensor

Computes the triplet loss between negative node pairs and sampled non-node pairs.

Parameters:

• z (Tensor) – The node embeddings.

• neg_edge_index (LongTensor) – The negative edge indices.