Adversarially Regularized Graph Autoencoder
By Shirui Pan, Ruiqi Hu, Guodong Long, Jing Jiang, Lina Yao, Chengqi Zhang. 2018
489 citations recorded by the source corpus.
Graph embedding is an effective method to represent graph data in a low dimensional space for graph analytics. Most existing embedding algorithms typically focus on preserving the topological structure or minimizing the reconstruction errors of graph data, but they have mostly ignored the data distribution of the latent codes from the graphs, which often results in inferior embedding in real-world graph data. In this paper, we propose a novel adversarial graph embedding framework for graph data. The framework encodes the topological structure and node content in a graph to a compact representation, on which a decoder is trained to reconstruct the graph structure. Furthermore, the latent representation is enforced to match a prior distribution via an adversarial training scheme. To learn a robust embedding, two variants of adversarial approaches, adversarially regularized graph autoencoder (ARGA) and adversarially regularized variational graph autoencoder (ARVGA), are developed. Experimental studies on real-world graphs validate our design and demonstrate that our algorithms outperform baselines by a wide margin in link prediction, graph clustering, and graph visualization tasks.
Subfields: Artificial Intelligence, Molecular Biology, Statistical and Nonlinear Physics, Representation & Self-Supervised Learning, Graph Learning
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
In Molecular Biology
In Statistical and Nonlinear Physics
In Representation & Self-Supervised Learning
In Graph Learning
Citations in the corpus
- Fast Graph Representation Learning with PyTorch Geometric (2019, background): “, 2019), autoencoding graphs (Kipf &Welling, 2016; Pan et al., 2018), aggregating jumping knowledge (Xu et al.”
- Self-Supervised Learning of Graph Neural Networks: A Unified Review (2021, methodology): “As the adversarial regularization is provably an equivalence of the JS-divergence between the distribution of the latent variable and the Gaussian prior, ARGA/ARVGA can achieve a similar effect to VGAE but with stronger regularization.”
- Network Representation Learning: A Survey (2017, unknown): “Adversarial Network Embedding (ANE) [125] and Adversarially Regularized Graph Autoencoder (ARGA) [126] learn robust vertex representations via enforcing an adversarial learning regularizer [58].”
- Graph Self-Supervised Learning: A Survey (2021, background): “ARGA/ARVGA [48] regularizes the GAE/VGAE model with generative adversarial networks (GANs) [55].”
- Attributed Graph Clustering via Adaptive Graph Convolution (2019, methodology): “In particular, GCN based methods such as GAE [ Kipf and Welling, 2016 ] , MGAE [ Wang et al. , 2017 ] , ARGE [ Pan et al. , 2018 ] have demonstrated state-of-the-art performance on several attributed graph clustering tasks.”
- GraphMAE: Self-Supervised Masked Graph Autoencoders (2022, background): “For GAEs [3, 20, 26, 27, 31] that leverage feature reconstruction, most of them still employ the vanilla architecture that risks learning trivial solutions.”
- How to Find Your Friendly Neighborhood: Graph Attention Design with Self-Supervision (2022, methodology): “Our DP attention to predict links is motivated by graph autoencoder (GAE) (Kipf & Welling, 2016) and its extensions (Pan et al., 2018; Park et al., 2019) reconstructing edges by applying a dot-product decoder to node representations.”
- Adaptive Graph Encoder for Attributed Graph Embedding (2020, methodology): “It is worth noting that the training objectives of GAE/VGAE and ARGA/ARVGA are the adjacency matrix reconstruction loss.”