Adversarial Network Embedding

By Quanyu Dai, Qiang Li, Jian Tang, Dan Wang. 2017

229 citations recorded by the source corpus.

Learning low-dimensional representations of networks has proved effective in a variety of tasks such as node classification, link prediction and network visualization. Existing methods can effectively encode different structural properties into the representations, such as neighborhood connectivity patterns, global structural role similarities and other high-order proximities. However, except for objectives to capture network structural properties, most of them suffer from lack of additional constraints for enhancing the robustness of representations. In this paper, we aim to exploit the strengths of generative adversarial networks in capturing latent features, and investigate its contribution in learning stable and robust graph representations. Specifically, we propose an Adversarial Network Embedding (ANE) framework, which leverages the adversarial learning principle to regularize the representation learning. It consists of two components, i.e., a structure preserving component and an adversarial learning component. The former component aims to capture network structural properties, while the latter contributes to learning robust representations by matching the posterior distribution of the latent representations to given priors. As shown by the empirical results, our method is competitive with or superior to state-of-the-art approaches on benchmark network embedding tasks.

Subfields: Artificial Intelligence, Cognitive Neuroscience, Molecular Biology

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 Cognitive Neuroscience

In Molecular Biology

Citations in the corpus

Attacking Graph-based Classification via Manipulating the Graph Structure (2019, background): “In particular, only a few studies [8, 15–17, 41, 57, 61, 77, 78] focused on adversarial graph-based machine learning.”
Adversarially Regularized Graph Autoencoder (2018, background): “However, their approach can only exploit the topological information [Dai et al., 2017].”
Self-Supervised Learning: Generative or Contrastive (2020, background): “ANE [23] Graph G-C AE Graph edges Link prediction - -”
Adversarial Attacks on Neural Networks for Graph Data (2018, background): “k) and especially we have dependencies between the nodes/data instances via the edges. 3 We want to clarify that this work is not about adversarial training (e.g. via GANs) for graph learning tasks [10] but about adversarial attacks. Adversarial Aacks on Classification Models for Graphs SIGKDD&apo”
Learning Graph Embedding With Adversarial Training Methods (2019, background): “Recently, there are some studies on applying the adversarial mechanism to graphs, such as AIDW [48] and NetRA [49].”
Semi-supervised Learning on Graphs with Generative Adversarial Nets (2018, background): “[7] leveraged adversarial learning to regularize training of graph representations.”
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].”
Adversarial Attacks on Node Embeddings via Graph Poisoning (2018, methodology): “Later, Dai et al. (2018a) proposed a reinforcement (Q-)learning formulation where they de-compose the selection of relevant edge ﬂips into selecting the two end-points.”
Deep Learning on Graphs: A Survey (2018, unknown): “Adversarial network embedding (ANE) [125] also adopted an adversarial training scheme to improve network embedding methods.”
Embedding Temporal Network via Neighborhood Formation (2018, background): “Network embedding has become a focal point of study in recent years, aiming at representing large-scale networks by mapping nodes to low-dimensional space [6, 9, 10, 20].”