torch_geometric.nn.pool.PANPooling

class PANPooling(in_channels: int, ratio: float = 0.5, min_score: Optional[float] = None, multiplier: float = 1.0, nonlinearity: Union[str, Callable] = 'tanh')[source]

Bases: Module

基于路径积分的池化操作符来自 “Path Integral Based Convolution and Pooling for Graph Neural Networks” 论文。

PAN池化执行顶部\(k\)池化，其中全局节点重要性基于节点特征和MET矩阵进行测量：

\[{\rm score} = \beta_1 \mathbf{X} \cdot \mathbf{p} + \beta_2 {\rm deg}(\mathbf{M})\]
Parameters:

  • in_channels (int) – Size of each input sample.

  • ratio (float) – 图池化比例，用于计算 \(k = \lceil \mathrm{ratio} \cdot N \rceil\). 如果 min_score 不为 None，则忽略此值。 (默认值: 0.5)

  • min_score (float, optional) – Minimal node score \(\tilde{\alpha}\) which is used to compute indices of pooled nodes \(\mathbf{i} = \mathbf{y}_i > \tilde{\alpha}\). When this value is not None, the ratio argument is ignored. (default: None)

  • multiplier (float, optional) – 池化后特征乘以的系数。这对于大型图和当使用min_score时非常有用。（默认值：1.0）

  • nonlinearity (str or callable, optional) – The non-linearity to use. (default: "tanh")

reset_parameters()[source]

重置模块的所有可学习参数。

forward(x: Tensor, M: SparseTensor, batch: Optional[Tensor] = None) → Tuple[Tensor, Tensor, Tensor, Optional[Tensor], Tensor, Tensor][source]

前向传播。

Parameters:

  • x (torch.Tensor) – The node feature matrix.

  • M (SparseTensor) – MET矩阵 \(\mathbf{M}\).

  • batch (torch.Tensor, optional) – The batch vector \(\mathbf{b} \in {\{ 0, \ldots, B-1\}}^N\), which assigns each node to a specific example. (default: None)

Return type:

Tuple[Tensor, Tensor, Tensor, Optional[Tensor], Tensor, Tensor]