def max_pool_graph_sage(x, edge_index, edge_weight,
self_kernel, neighbor_mlp_kernel, neighbor_kernel,
neighbor_mlp_bias=None, bias=None, activation=None,
concat=True, normalize=False):
"""
:param x: Tensor, shape: [num_nodes, num_features], node features
:param edge_index: Tensor, shape: [2, num_edges], edge information
:param edge_weight: Tensor or None, shape: [num_edges]
:param self_kernel: Tensor, shape: [num_features, num_hidden_units], weight.
:param neighbor_mlp_kernel: Tensor, shape: [num_features, num_hidden_units]. weight.
:param neighbor_kernel: Tensor, shape: [num_hidden_units, num_hidden_units], weight.
:param neighbor_mlp_bias: Tensor, shape: [num_hidden_units * 2], bias
:param bias: Tensor, shape: [num_output_features], bias.
:param activation: Activation function to use.
:param normalize: If set to :obj:`True`, output features
will be :math:`\ell_2`-normalized, *i.e.*,
:math:`\frac{\mathbf{x}^{\prime}_i}
{\| \mathbf{x}^{\prime}_i \|_2}`.
(default: :obj:`False`)
:return: Updated node features (x), shape: [num_nodes, num_output_features]
"""
num_nodes = tf.shape(x)[0]
num_edges = tf.shape(edge_index)[1]
if edge_weight is not None:
edge_weight = tf.ones([num_edges], dtype=tf.float32)
row, col = edge_index[0], edge_index[1]
repeated_x = tf.gather(x, row)
neighbor_x = tf.gather(x, col)
neighbor_x = gcn_mapper(repeated_x, neighbor_x, edge_weight=edge_weight)
h = neighbor_x @ neighbor_mlp_kernel
if neighbor_mlp_bias is not None:
h += neighbor_mlp_bias
if activation is not None:
h = activation(h)
reduced_h = max_reducer(h, row, num_nodes=num_nodes)
from_neighs = reduced_h @ neighbor_kernel
from_x = x @ self_kernel
if concat:
output = tf.concat([from_x, from_neighs], axis=1)
else:
output = from_x + from_neighs
if bias is not None:
output += bias
if activation is not None:
output = activation(output)
if normalize:
output = tf.nn.l2_normalize(output, axis=-1)
return output
def mean_graph_sage(x, edge_index, edge_weight,
self_kernel,
neighbor_kernel,
bias=None,
activation=None,
concat=True, normalize=False):
"""
:param x: Tensor, shape: [num_nodes, num_features], node features
:param edge_index: Tensor, shape: [2, num_edges], edge information
:param edge_weight: Tensor or None, shape: [num_edges]
:param self_kernel: Tensor, shape: [num_features, num_hidden_units], weight
:param neighbor_kernel: Tensor, shape: [num_features, num_hidden_units], weight.
:param bias: Tensor, shape: [num_output_features], bias
:param activation: Activation function to use.
:param normalize: If set to :obj:`True`, output features
will be :math:`\ell_2`-normalized, *i.e.*,
:math:`\frac{\mathbf{x}^{\prime}_i}
{\| \mathbf{x}^{\prime}_i \|_2}`.
(default: :obj:`False`)
:return: Updated node features (x), shape: [num_nodes, num_output_features]
"""
num_nodes = tf.shape(x)[0]
# num_edges = tf.shape(edge_index)[1]
row, col = edge_index[0], edge_index[1]
repeated_x = tf.gather(x, row)
neighbor_x = tf.gather(x, col)
if edge_weight is not None:
neighbor_x = gcn_mapper(repeated_x, neighbor_x, edge_weight=edge_weight)
neighbor_reduced_msg = mean_reducer(neighbor_x, row, num_nodes=num_nodes)
neighbor_msg = neighbor_reduced_msg @ neighbor_kernel
x = x @ self_kernel
if concat:
h = tf.concat([x, neighbor_msg], axis=1)
else:
h = x + neighbor_msg
if bias is not None:
h += bias
if activation is not None:
h = activation(h)
if normalize:
h = tf.nn.l2_normalize(h, axis=-1)
return h
def gcn_graph_sage(x,
edge_index,
edge_weight,
kernel,
bias=None,
activation=None,
normalize=False,
cache=None):
"""
:param x: Tensor, shape: [num_nodes, num_features], node features
:param edge_index: Tensor, shape: [2, num_edges], edge information
:param edge_weight: Tensor or None, shape: [num_edges]
:param kernel: Tensor, shape: [num_features, num_output_features], weight
:param bias: Tensor, shape: [num_output_features], bias
:param activation: Activation function to use.
:param normalize: If set to :obj:`True`, output features
will be :math:`\ell_2`-normalized, *i.e.*,
:math:`\frac{\mathbf{x}^{\prime}_i}
{\| \mathbf{x}^{\prime}_i \|_2}`.
(default: :obj:`False`)
:param cache: A dict for caching A' for GCN. Different graph should not share the same cache dict.
:return: Updated node features (x), shape: [num_nodes, num_output_features]
"""
if edge_weight is not None:
edge_weight = tf.ones([edge_index.shape[1]], dtype=tf.float32)
updated_edge_index, normed_edge_weight = gcn_norm_edge(
edge_index, x.shape[0], edge_weight, cache)
row, col = updated_edge_index
repeated_x = tf.gather(x, row)
neighbor_x = tf.gather(x, col)
neighbor_x = gcn_mapper(repeated_x,
neighbor_x,
edge_weight=normed_edge_weight)
reduced_msg = sum_reducer(neighbor_x, row, num_nodes=x.shape[0])
h = reduced_msg @ kernel
if bias is not None:
h += bias
if activation is not None:
h = activation(h)
if normalize:
h = tf.nn.l2_normalize(h, axis=-1)
return h