/examples/pytorch/gas/model.py

import torch.nn as nn
import dgl.function as fn
import torch as th
from dgl.nn.functional import edge_softmax


class MLP(nn.Module):
    def __init__(self, in_dim, out_dim):
        super().__init__()
        self.W = nn.Linear(in_dim, out_dim)

    def apply_edges(self, edges):
        h_e = edges.data['h']
        h_u = edges.src['h']
        h_v = edges.dst['h']
        score = self.W(th.cat([h_e, h_u, h_v], -1))
        return {'score': score}

    def forward(self, g, e_feat, u_feat, v_feat):
        with g.local_scope():
            g.edges['forward'].data['h'] = e_feat
            g.nodes['u'].data['h'] = u_feat
            g.nodes['v'].data['h'] = v_feat
            g.apply_edges(self.apply_edges, etype="forward")
            return g.edges['forward'].data['score']


class GASConv(nn.Module):
    """One layer of GAS."""

    def __init__(self,
                 e_in_dim,
                 u_in_dim,
                 v_in_dim,
                 e_out_dim,
                 u_out_dim,
                 v_out_dim,
                 activation=None,
                 dropout=0):
        super(GASConv, self).__init__()

        self.activation = activation
        self.dropout = nn.Dropout(dropout)

        self.e_linear = nn.Linear(e_in_dim, e_out_dim)
        self.u_linear = nn.Linear(u_in_dim, e_out_dim)
        self.v_linear = nn.Linear(v_in_dim, e_out_dim)

        self.W_ATTN_u = nn.Linear(u_in_dim, v_in_dim + e_in_dim)
        self.W_ATTN_v = nn.Linear(v_in_dim, u_in_dim + e_in_dim)

        # the proportion of h_u and h_Nu are specified as 1/2 in formula 8
        nu_dim = int(u_out_dim / 2)
        nv_dim = int(v_out_dim / 2)

        self.W_u = nn.Linear(v_in_dim + e_in_dim, nu_dim)
        self.W_v = nn.Linear(u_in_dim + e_in_dim, nv_dim)

        self.Vu = nn.Linear(u_in_dim, u_out_dim - nu_dim)
        self.Vv = nn.Linear(v_in_dim, v_out_dim - nv_dim)

    def forward(self, g, e_feat, u_feat, v_feat):
        with g.local_scope():
            g.nodes['u'].data['h'] = u_feat
            g.nodes['v'].data['h'] = v_feat
            g.edges['forward'].data['h'] = e_feat
            g.edges['backward'].data['h'] = e_feat

            # formula 3 and 4 (optimized implementation to save memory)
            g.nodes["u"].data.update({'he_u': self.u_linear(u_feat)})
            g.nodes["v"].data.update({'he_v': self.v_linear(v_feat)})
            g.edges["forward"].data.update({'he_e': self.e_linear(e_feat)})
            g.apply_edges(lambda edges: {'he': edges.data['he_e'] + edges.src['he_u'] + edges.dst['he_v']}, etype='forward')
            he = g.edges["forward"].data['he']
            if self.activation is not None:
                he = self.activation(he)

            # formula 6
            g.apply_edges(lambda edges: {'h_ve': th.cat([edges.src['h'], edges.data['h']], -1)}, etype='backward')
            g.apply_edges(lambda edges: {'h_ue': th.cat([edges.src['h'], edges.data['h']], -1)}, etype='forward')

            # formula 7, self-attention
            g.nodes['u'].data['h_att_u'] = self.W_ATTN_u(u_feat)
            g.nodes['v'].data['h_att_v'] = self.W_ATTN_v(v_feat)

            # Step 1: dot product
            g.apply_edges(fn.e_dot_v('h_ve', 'h_att_u', 'edotv'), etype='backward')
            g.apply_edges(fn.e_dot_v('h_ue', 'h_att_v', 'edotv'), etype='forward')

            # Step 2. softmax
            g.edges['backward'].data['sfm'] = edge_softmax(g['backward'], g.edges['backward'].data['edotv'])
            g.edges['forward'].data['sfm'] = edge_softmax(g['forward'], g.edges['forward'].data['edotv'])

            # Step 3. Broadcast softmax value to each edge, and then attention is done
            g.apply_edges(lambda edges: {'attn': edges.data['h_ve'] * edges.data['sfm']}, etype='backward')
            g.apply_edges(lambda edges: {'attn': edges.data['h_ue'] * edges.data['sfm']}, etype='forward')

            # Step 4. Aggregate attention to dst,user nodes, so formula 7 is done
            g.update_all(fn.copy_e('attn', 'm'), fn.sum('m', 'agg_u'), etype='backward')
            g.update_all(fn.copy_e('attn', 'm'), fn.sum('m', 'agg_v'), etype='forward')

            # formula 5
            h_nu = self.W_u(g.nodes['u'].data['agg_u'])
            h_nv = self.W_v(g.nodes['v'].data['agg_v'])
            if self.activation is not None:
                h_nu = self.activation(h_nu)
                h_nv = self.activation(h_nv)

            # Dropout
            he = self.dropout(he)
            h_nu = self.dropout(h_nu)
            h_nv = self.dropout(h_nv)

            # formula 8
            hu = th.cat([self.Vu(u_feat), h_nu], -1)
            hv = th.cat([self.Vv(v_feat), h_nv], -1)

            return he, hu, hv


class GAS(nn.Module):
    def __init__(self,
                 e_in_dim,
                 u_in_dim,
                 v_in_dim,
                 e_hid_dim,
                 u_hid_dim,
                 v_hid_dim,
                 out_dim,
                 num_layers=2,
                 dropout=0.0,
                 activation=None):
        super(GAS, self).__init__()
        self.e_in_dim = e_in_dim
        self.u_in_dim = u_in_dim
        self.v_in_dim = v_in_dim
        self.e_hid_dim = e_hid_dim
        self.u_hid_dim = u_hid_dim
        self.v_hid_dim = v_hid_dim
        self.out_dim = out_dim
        self.num_layer = num_layers
        self.dropout = dropout
        self.activation = activation
        self.predictor = MLP(e_hid_dim + u_hid_dim + v_hid_dim, out_dim)
        self.layers = nn.ModuleList()

        # Input layer
        self.layers.append(GASConv(self.e_in_dim,
                                   self.u_in_dim,
                                   self.v_in_dim,
                                   self.e_hid_dim,
                                   self.u_hid_dim,
                                   self.v_hid_dim,
                                   activation=self.activation,
                                   dropout=self.dropout))

        # Hidden layers with n - 1 CompGraphConv layers
        for i in range(self.num_layer - 1):
            self.layers.append(GASConv(self.e_hid_dim,
                                       self.u_hid_dim,
                                       self.v_hid_dim,
                                       self.e_hid_dim,
                                       self.u_hid_dim,
                                       self.v_hid_dim,
                                       activation=self.activation,
                                       dropout=self.dropout))

    def forward(self, graph, e_feat, u_feat, v_feat):
        # For full graph training, directly use the graph
        # Forward of n layers of GAS
        for layer in self.layers:
            e_feat, u_feat, v_feat = layer(graph, e_feat, u_feat, v_feat)

        # return the result of final prediction layer
        return self.predictor(graph, e_feat, u_feat, v_feat)