Adversarially Regularized Graph Autoencoder for Graph Embedding

Pan, Shirui; Hu, Ruiqi; Long, Guodong; Jiang, Jing; Yao, Lina; Zhang, Chengqi

doi:10.48550/arxiv.1802.04407

Cited by 50 publications

(59 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1-A). Drawing inspiration from ARVGA [13], our framework is composed of a variational autoencoder A align and a discriminator D align . The A align comprises a probabilistic encoder that encodes an input as a distribution over the latent space instead of encoding an input as a single point.…”

Section: Proposed Methodsmentioning

confidence: 99%

“…To address the challenges above and motivated by the recent development of graph neural network-based solutions, we propose a multi-resolution Stairway-GraphNet (SG-Net) method to jointly predict and super-resolve a target graph modality based on a given modality in both inter-and intra-domains. To do so, prior to the prediction blocks, we propose an inter-modality aligner network based on adversarially regularized variational graph autoencoder (ARVGA) [13] to align the training graphs of the source modality to that of the target one. Second, given the aligned source graphs, we design an inter-modality superresolution graph GAN (gGAN) to map the aligned source graph from one modality (e.g., morphological) to the target modality (e.g., functional).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

StairwayGraphNet for Inter- and Intra-modality Multi-resolution Brain Graph Alignment and Synthesis

Mhiri¹,

Mahjoub²,

Rekik³

2021

Preprint

View full text Add to dashboard Cite

Synthesizing multimodality medical data provides complementary knowledge and helps doctors make precise clinical decisions. Although promising, existing multimodal brain graph synthesis frameworks have several limitations. First, they mainly tackle only one problem (intra-or inter-modality), limiting their generalizability to synthesizing inter-and intra-modality simultaneously. Second, while few techniques work on super-resolving low-resolution brain graphs within a single modality (i.e., intra), inter-modality graph super-resolution remains unexplored though this would avoid the need for costly data collection and processing. More importantly, both target and source domains might have different distributions, which causes a domain fracture between them. To fill these gaps, we propose a multi-resolution StairwayGraphNet (SG-Net) framework to jointly infer a target graph modality based on a given modality and super-resolve brain graphs in both inter and intra domains. Our SG-Net is grounded in three main contributions: (i) predicting a target graph from a source one based on a novel graph generative adversarial network in both inter (e.g., morphological-functional) and intra (e.g., functional-functional) domains, (ii) generating high-resolution brain graphs without resorting to the time consuming and expensive MRI processing steps, and (iii) enforcing the source distribution to match that of the ground truth graphs using an inter-modality aligner to relax the loss function to optimize. Moreover, we design a new Ground Truth-Preserving loss function to guide both generators in learning the topological structure of ground truth brain graphs more accurately. Our comprehensive experiments on predicting target brain graphs from source graphs using a multi-resolution stairway showed the outperformance of our method in comparison with its variants and state-of-the-art method. SG-Net presents the first work for graph alignment and synthesis across varying modalities and resolutions, which handles graph size, distribution, and structure variations. Our Python SG-Net code is available on BASIRA GitHub at https://github.com/basiralab/SG-Net.

show abstract

Section: Proposed Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

StairwayGraphNet for Inter- and Intra-modality Multi-resolution Brain Graph Alignment and Synthesis

Mhiri¹,

Mahjoub²,

Rekik³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Metrics. Following [12,30], we use accuracy (ACC), normalized mutual information (NMI), precision, F-score(F1) and average rand index (ARI) as our evaluation metrics. We conduct 10 repeat experiments.…”

Section: Node Clusteringmentioning

confidence: 99%

On Generalization of Graph Autoencoders with Adversarial Training

Huang¹,

Pei²,

Menkovski³

et al. 2021

Preprint

View full text Add to dashboard Cite

Adversarial training is an approach for increasing model's resilience against adversarial perturbations. Such approaches have been demonstrated to result in models with feature representations that generalize better. However, limited works have been done on adversarial training of models on graph data. In this paper, we raise such a question -does adversarial training improve the generalization of graph representations. We formulate L2 and L∞ versions of adversarial training in two powerful node embedding methods: graph autoencoder (GAE) and variational graph autoencoder (VGAE). We conduct extensive experiments on three main applications, i.e. link prediction, node clustering, graph anomaly detection of GAE and VGAE, and demonstrate that both L2 and L∞ adversarial training boost the generalization of GAE and VGAE.

show abstract

“…Also, graph autoencoder (GAE) and varitional graph auto-encoder (VGAE) have been proposed which using the GCN as encoder and the inner product between node embeddings as decoder in auto-encoder or variational auto-encoder framework for unsupervised representation learning [18], [30]. Based on GAE and VGAE, Pan et al [15] have proposed ARGA and ARVGA with adversarial training scheme where the latent node representations are enforced to match a prior distribution. Wang et al [16] have proposed MGAE which is a GAE with a marginalization process to disturb the network attribute information.…”

Section: Deep Learning For Network Clusteringmentioning

confidence: 99%

“…To handle the challenges, many network embedding and graph neural network related methods have been developed recently for node representation learning to improve the accuracy of downstream applications like graph classification, link prediction and graph clustering. The representative methods include DeepWalk [9], Line [10], node2vec [11], struc2vec [12], GCN [13], GraphSAGE [14], ARGA/ARVGA [15], MGAE [16], AGC [8], etc. For example, DeepWalk [9] and Node2Vec [11] are two representative network embedding methods to learn low-dimensional representations for nodes through the local neighborhood structure prediction.…”

Section: Introductionmentioning

confidence: 99%

Variational Co-embedding Learning for Attributed Network Clustering

Yang,

Verma,

Cai

et al. 2021

Preprint

View full text Add to dashboard Cite

Recent works for attributed network clustering utilize graph convolution to obtain node embeddings and simultaneously perform clustering assignments on the embedding space. It is effective since graph convolution combines the structural and attributive information for node embedding learning. However, a major limitation of such works is that the graph convolution only incorporates the attribute information from the local neighborhood of nodes but fails to exploit the mutual affinities between nodes and attributes. In this regard, we propose a variational co-embedding learning model for attributed network clustering (VCLANC). VCLANC is composed of dual variational auto-encoders to simultaneously embed nodes and attributes. Relying on this, the mutual affinity information between nodes and attributes could be reconstructed from the embedding space and served as extra self-supervised knowledge for representation learning. At the same time, trainable Gaussian mixture model is used as priors to infer the node clustering assignments. To strengthen the performance of the inferred clusters, we use a mutual distance loss on the centers of the Gaussian priors and a clustering assignment hardening loss on the node embeddings. Experimental results on four real-world attributed network datasets demonstrate the effectiveness of the proposed VCLANC for attributed network clustering.

show abstract

Adversarially Regularized Graph Autoencoder for Graph Embedding

Cited by 50 publications

References 14 publications

StairwayGraphNet for Inter- and Intra-modality Multi-resolution Brain Graph Alignment and Synthesis

StairwayGraphNet for Inter- and Intra-modality Multi-resolution Brain Graph Alignment and Synthesis

On Generalization of Graph Autoencoders with Adversarial Training

Variational Co-embedding Learning for Attributed Network Clustering

Contact Info

Product

Resources

About