Implement NGCF (PreferredAI#529)

* Generated model base from LightGCN

* wip

* wip example

* add self-connection

* refactor code

* added sanity check

* Changed train batch size in example to 1024

* Updated readme for example folder

* Update Readme

* update docs

* Update block comment

---------

Co-authored-by: tqtg <tuantq.vnu@gmail.com>

README.md

@@ -117,6 +117,7 @@ The recommender models supported by Cornac are listed below. Why don't you join
 |      | [Hybrid neural recommendation with joint deep representation learning of ratings and reviews (HRDR)](cornac/models/hrdr), [paper](https://www.sciencedirect.com/science/article/abs/pii/S0925231219313207) | [requirements.txt](cornac/models/hrdr/requirements.txt) | [hrdr_example.py](examples/hrdr_example.py)
 |      | [LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation](cornac/models/lightgcn), [paper](https://arxiv.org/pdf/2002.02126.pdf) | [requirements.txt](cornac/models/lightgcn/requirements.txt) | [lightgcn_example.py](examples/lightgcn_example.py)
 | 2019 | [Embarrassingly Shallow Autoencoders for Sparse Data (EASEᴿ)](cornac/models/ease), [paper](https://arxiv.org/pdf/1905.03375.pdf) | N/A | [ease_movielens.py](examples/ease_movielens.py)
+|      | [Neural Graph Collaborative Filtering](cornac/models/ngcf), [paper](https://arxiv.org/pdf/1905.08108.pdf) | [requirements.txt](cornac/models/ngcf/requirements.txt) | [ngcf_example.py](examples/ngcf_example.py)
 | 2018 | [Collaborative Context Poisson Factorization (C2PF)](cornac/models/c2pf), [paper](https://www.ijcai.org/proceedings/2018/0370.pdf) | N/A | [c2pf_exp.py](examples/c2pf_example.py)
 |      | [Graph Convolutional Matrix Completion (GCMC)](cornac/models/gcmc), [paper](https://www.kdd.org/kdd2018/files/deep-learning-day/DLDay18_paper_32.pdf) | [requirements.txt](cornac/models/gcmc/requirements.txt) | [gcmc_example.py](examples/gcmc_example.py)
 |      | [Multi-Task Explainable Recommendation (MTER)](cornac/models/mter), [paper](https://arxiv.org/pdf/1806.03568.pdf) | N/A | [mter_exp.py](examples/mter_example.py)

cornac/models/__init__.py

@@ -51,6 +51,7 @@
 from .ncf import GMF
 from .ncf import MLP
 from .ncf import NeuMF
+from .ngcf import NGCF
 from .nmf import NMF
 from .online_ibpr import OnlineIBPR
 from .pcrl import PCRL

cornac/models/ngcf/__init__.py

@@ -0,0 +1,16 @@
+# Copyright 2018 The Cornac Authors. All Rights Reserved.
+#
+# 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.
+# ============================================================================
+
+from .recom_ngcf import NGCF

cornac/models/ngcf/ngcf.py

@@ -0,0 +1,185 @@
+# Reference: https://github.com/dmlc/dgl/blob/master/examples/pytorch/NGCF/NGCF/model.py
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import dgl
+import dgl.function as fn
+
+
+USER_KEY = "user"
+ITEM_KEY = "item"
+
+
+def construct_graph(data_set):
+    """
+    Generates graph given a cornac data set
+
+    Parameters
+    ----------
+    data_set : cornac.data.dataset.Dataset
+        The data set as provided by cornac
+    """
+    user_indices, item_indices, _ = data_set.uir_tuple
+
+    # construct graph from the train data and add self-loops
+    user_selfs = [i for i in range(data_set.total_users)]
+    item_selfs = [i for i in range(data_set.total_items)]
+
+    data_dict = {
+        (USER_KEY, "user_self", USER_KEY): (user_selfs, user_selfs),
+        (ITEM_KEY, "item_self", ITEM_KEY): (item_selfs, item_selfs),
+        (USER_KEY, "user_item", ITEM_KEY): (user_indices, item_indices),
+        (ITEM_KEY, "item_user", USER_KEY): (item_indices, user_indices),
+    }
+    num_dict = {USER_KEY: data_set.total_users, ITEM_KEY: data_set.total_items}
+
+    return dgl.heterograph(data_dict, num_nodes_dict=num_dict)
+
+
+class NGCFLayer(nn.Module):
+    def __init__(self, in_size, out_size, norm_dict, dropout):
+        super(NGCFLayer, self).__init__()
+        self.in_size = in_size
+        self.out_size = out_size
+
+        # weights for different types of messages
+        self.W1 = nn.Linear(in_size, out_size, bias=True)
+        self.W2 = nn.Linear(in_size, out_size, bias=True)
+
+        # leaky relu
+        self.leaky_relu = nn.LeakyReLU(0.2)
+
+        # dropout layer
+        self.dropout = nn.Dropout(dropout)
+
+        # initialization
+        torch.nn.init.xavier_uniform_(self.W1.weight)
+        torch.nn.init.constant_(self.W1.bias, 0)
+        torch.nn.init.xavier_uniform_(self.W2.weight)
+        torch.nn.init.constant_(self.W2.bias, 0)
+
+        # norm
+        self.norm_dict = norm_dict
+
+    def forward(self, g, feat_dict):
+        funcs = {}  # message and reduce functions dict
+        # for each type of edges, compute messages and reduce them all
+        for srctype, etype, dsttype in g.canonical_etypes:
+            if srctype == dsttype:  # for self loops
+                messages = self.W1(feat_dict[srctype])
+                g.nodes[srctype].data[etype] = messages  # store in ndata
+                funcs[(srctype, etype, dsttype)] = (
+                    fn.copy_u(etype, "m"),
+                    fn.sum("m", "h"),
+                )  # define message and reduce functions
+            else:
+                src, dst = g.edges(etype=(srctype, etype, dsttype))
+                norm = self.norm_dict[(srctype, etype, dsttype)]
+                messages = norm * (
+                    self.W1(feat_dict[srctype][src])
+                    + self.W2(feat_dict[srctype][src] * feat_dict[dsttype][dst])
+                )  # compute messages
+                g.edges[(srctype, etype, dsttype)].data[
+                    etype
+                ] = messages  # store in edata
+                funcs[(srctype, etype, dsttype)] = (
+                    fn.copy_e(etype, "m"),
+                    fn.sum("m", "h"),
+                )  # define message and reduce functions
+
+        g.multi_update_all(
+            funcs, "sum"
+        )  # update all, reduce by first type-wisely then across different types
+        feature_dict = {}
+        for ntype in g.ntypes:
+            h = self.leaky_relu(g.nodes[ntype].data["h"])  # leaky relu
+            h = self.dropout(h)  # dropout
+            h = F.normalize(h, dim=1, p=2)  # l2 normalize
+            feature_dict[ntype] = h
+        return feature_dict
+
+
+class Model(nn.Module):
+    def __init__(self, g, in_size, layer_sizes, dropout_rates, lambda_reg, device=None):
+        super(Model, self).__init__()
+        self.norm_dict = dict()
+        self.lambda_reg = lambda_reg
+        self.device = device
+
+        for srctype, etype, dsttype in g.canonical_etypes:
+            src, dst = g.edges(etype=(srctype, etype, dsttype))
+            dst_degree = g.in_degrees(
+                dst, etype=(srctype, etype, dsttype)
+            ).float()  # obtain degrees
+            src_degree = g.out_degrees(src, etype=(srctype, etype, dsttype)).float()
+            norm = torch.pow(src_degree * dst_degree, -0.5).unsqueeze(1)  # compute norm
+            self.norm_dict[(srctype, etype, dsttype)] = norm
+
+        self.layers = nn.ModuleList()
+
+        # sanity check, just to ensure layer sizes and dropout_rates have the same size
+        assert len(layer_sizes) == len(dropout_rates), "'layer_sizes' and " \
+            "'dropout_rates' must be of the same size"
+
+        self.layers.append(
+            NGCFLayer(in_size, layer_sizes[0], self.norm_dict, dropout_rates[0])
+        )
+        self.num_layers = len(layer_sizes)
+        for i in range(self.num_layers - 1):
+            self.layers.append(
+                NGCFLayer(
+                    layer_sizes[i],
+                    layer_sizes[i + 1],
+                    self.norm_dict,
+                    dropout_rates[i + 1],
+                )
+            )
+        self.initializer = nn.init.xavier_uniform_
+
+        # embeddings for different types of nodes
+        self.feature_dict = nn.ParameterDict(
+            {
+                ntype: nn.Parameter(
+                    self.initializer(torch.empty(g.num_nodes(ntype), in_size))
+                )
+                for ntype in g.ntypes
+            }
+        )
+
+    def forward(self, g, users=None, pos_items=None, neg_items=None):
+        h_dict = {ntype: self.feature_dict[ntype] for ntype in g.ntypes}
+        # obtain features of each layer and concatenate them all
+        user_embeds = []
+        item_embeds = []
+        user_embeds.append(h_dict[USER_KEY])
+        item_embeds.append(h_dict[ITEM_KEY])
+        for layer in self.layers:
+            h_dict = layer(g, h_dict)
+            user_embeds.append(h_dict[USER_KEY])
+            item_embeds.append(h_dict[ITEM_KEY])
+        user_embd = torch.cat(user_embeds, 1)
+        item_embd = torch.cat(item_embeds, 1)
+
+        u_g_embeddings = user_embd if users is None else user_embd[users, :]
+        pos_i_g_embeddings = item_embd if pos_items is None else item_embd[pos_items, :]
+        neg_i_g_embeddings = item_embd if neg_items is None else item_embd[neg_items, :]
+
+        return u_g_embeddings, pos_i_g_embeddings, neg_i_g_embeddings
+
+    def loss_fn(self, users, pos_items, neg_items):
+        pos_scores = (users * pos_items).sum(1)
+        neg_scores = (users * neg_items).sum(1)
+
+        bpr_loss = F.softplus(neg_scores - pos_scores).mean()
+        reg_loss = (
+            (1 / 2)
+            * (
+                torch.norm(users) ** 2
+                + torch.norm(pos_items) ** 2
+                + torch.norm(neg_items) ** 2
+            )
+            / len(users)
+        )
+
+        return bpr_loss + self.lambda_reg * reg_loss, bpr_loss, reg_loss