Learning Graph Representations with Embedding Propagation

By Alberto García-Durán, Mathias Niepert. 2017

176 citations recorded by the source corpus.

We propose Embedding Propagation (EP), an unsupervised learning framework for graph-structured data. EP learns vector representations of graphs by passing two types of messages between neighboring nodes. Forward messages consist of label representations such as representations of words and other attributes associated with the nodes. Backward messages consist of gradients that result from aggregating the label representations and applying a reconstruction loss. Node representations are finally computed from the representation of their labels. With significantly fewer parameters and hyperparameters an instance of EP is competitive with and often outperforms state of the art unsupervised and semi-supervised learning methods on a range of benchmark data sets.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Relational inductive biases, deep learning, and graph networks (2018, background): “…work has also focused on building generative models of graphs (Li et al., 2018; De Cao and Kipf, 2018; You et al., 2018; Bojchevski et al., 2018), and unsupervised learning of graph embeddings (Perozzi et al., 2014; Tang et al., 2015; Grover and Leskovec, 2016; Garćıa-Durán and Niepert, 2017).”
Deep Graph Infomax (2018, methodology): “…representation learning with graph-structured data rely on random walk-based objectives (Grover & Leskovec, 2016; Perozzi et al., 2014; Tang et al., 2015; Hamilton et al., 2017a), sometimes further simpliﬁed to reconstruct adjacency information (Kipf & Welling, 2016b; Duran & Niepert, 2017).”
Graph Representation Learning via Graphical Mutual Information Maximization (2020, background): “Here, results for EP-B [11], DGI [41], Planetoid-T [45], GAT [40], as”
LanczosNet: Multi-Scale Deep Graph Convolutional Networks (2019, background): “For the latter, unsupervised node/graph embedding learning [22, 23] is common.”
AliGraph: A Comprehensive Graph Neural Network Platform (2019, methodology): “This algorithm is designed to mitigate the heavy computation and storage costs of the traditional embedding propagation (EP) algorithm [12] on heterogeneous networks, HEP [56].”
Link Prediction Based on Graph Neural Networks (2018, methodology): “Network embedding methods [19, 21, 20, 39, 40, 41] have gained great popularity recently.”
Contrastive Multi-View Representation Learning on Graphs (2020, background): “Graph autoencoders (GAE) (Kipf & Welling, 2016; Garcia Duran & Niepert, 2017; Wang et al., 2017; Pan et al., 2018; Park et al., 2019) train encoders that impose the topological closeness of nodes in the graph structure on the latent space by predicting the ﬁrst-order neighbors.”
Learning Discrete Structures for Graph Neural Networks (2019, background): “…of graph convolutional networks (GCN), a class of GNNs. Intuitively, GCNs learn node representations by passing and aggregating messages between neighboring nodes (Kipf & Welling, 2017; Monti et al., 2017; Gilmer et al., 2017; Hamilton et al., 2017; Duran & Niepert, 2017; Velickovic et al., 2018).”
Neural Graph Collaborative Filtering (2019, methodology): “Next we build upon the message-passing architecture of GNNs [8, 37] in order to capture CF signal along the graph structure and refine the embeddings of users and items.”
Bootstrapped Representation Learning on Graphs (2021, methodology): “Earlier combinations of GNNs and self-supervised learning involve Embedding Propagation (Garc´ıa-Dur´an & Niepert, 2017), Variational Graph Autoencoders (Kipf & Welling, 2016) and Graph2Gauss (Bojchevski & G¨unnemann, 2018).”