# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""AGNNConv Layer."""
import mindspore as ms
from mindspore_gl import Graph
from .. import GNNCell
[文档]class AGNNConv(GNNCell):
r"""
Attention Based Graph Neural Network.
From the paper `Attention-based Graph Neural Network for Semi-Supervised Learning <https://arxiv.org/abs/1803.03735>`_ .
.. math::
H^{l+1} = P H^{l}
Computation of :math:`P` is:
.. math::
P_{ij} = \mathrm{softmax}_i ( \beta \cdot \cos(h_i^l, h_j^l))
:math:`\beta` is a single scalar parameter.
Args:
init_beta (float, optional): Init :math:`\beta`, a single scalar parameter. Default: ``1.0``.
learn_beta (bool, optional): Whether :math:`\beta` is learnable. Default: ``True``.
Inputs:
- **x** (Tensor): The input node features. The shape is :math:`(N,*)` where :math:`N` is the number of nodes,
and :math:`*` could be of any shape.
- **g** (Graph): The input graph.
Outputs:
- Tensor, output node features, where the shape should be the same as input 'x'.
Raises:
TypeError: If `init_beta` is not a float.
TypeError: If `learn_beta` is not a bool.
Supported Platforms:
``Ascend`` ``GPU``
Examples:
>>> import mindspore as ms
>>> from mindspore_gl.nn import AGNNConv
>>> from mindspore_gl import GraphField
>>> n_nodes = 4
>>> n_edges = 8
>>> feat_size = 16
>>> src_idx = ms.Tensor([0, 0, 0, 1, 1, 1, 2, 3], ms.int32)
>>> dst_idx = ms.Tensor([0, 1, 3, 1, 2, 3, 3, 2], ms.int32)
>>> ones = ms.ops.Ones()
>>> feat = ones((n_nodes, feat_size), ms.float32)
>>> graph_field = GraphField(src_idx, dst_idx, n_nodes, n_edges)
>>> conv = AGNNConv()
>>> ret = conv(feat, *graph_field.get_graph())
>>> print(ret.shape)
(4, 16)
"""
def __init__(self,
init_beta: float = 1.0,
learn_beta: bool = True):
super().__init__()
if not isinstance(init_beta, float):
raise ValueError("init_beta must be float")
if not isinstance(learn_beta, bool):
raise ValueError("learn_beta must be bool")
if learn_beta:
self.beta = ms.Parameter(ms.Tensor([init_beta], ms.float32))
else:
self.beta = ms.Tensor([init_beta], ms.float32)
# pylint: disable=arguments-differ
def construct(self, x, g: Graph):
"""
Construct function for AGNNConv.
"""
g.set_vertex_attr({"h": x, "norm_h": ms.ops.L2Normalize()(x)})
for v in g.dst_vertex:
cosine_dis = [ms.ops.Exp()(self.beta * g.dot(u.norm_h, v.norm_h)) for u in v.innbs]
a = cosine_dis / g.sum(cosine_dis)
v.h = g.sum([u.h for u in v.innbs] * a)
return [v.h for v in g.dst_vertex]