Elementary Subgraph Features for Link Prediction With Neural Networks

Fang, Zhihong; Tan, Shaolin; Wang, Yaonan; Lü, Jinhu

doi:10.1109/tkde.2021.3132352

Cited by 8 publications

(3 citation statements)

References 58 publications

(85 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The subgraph-based approach involves learning features from subgraphs to predict connections. Fang et al [4] proposed an architecture for link prediction called Neural Networks with Elementary Subgraphs Features (NNESF), which has relatively low computational complexity and a small number of hyperparameters. Lai et al proposed the ARCLink model [48], which is capable of creating a more efficient subgraph vector representation.…”

Section: Link Predictionmentioning

confidence: 99%

See 1 more Smart Citation

Rcoco: contrastive collective link prediction across multiplex network in Riemannian space

Sun,

Li,

Yang

et al. 2024

Int. J. Mach. Learn. & Cyber.

View full text Add to dashboard Cite

Link prediction typically studies the probability of future interconnection among nodes with the observation in a single social network. More often than not, real scenario is presented as a multiplex network with common (anchor) users active in multiple social networks. In the literature, most existing works study either the intra-link prediction in a single network or inter-link prediction among networks (a.k.a. network alignment), and consider two learning tasks are independent from each other, which is still away from the fact. On the representation space, the vast majority of existing methods are built upon the traditional Euclidean space, unaware of the inherent geometry of social networks. The third issue is on the scarce anchor users. Annotating anchor users is laborious and expensive, and thus it is impractical to work with quantities of anchor users. Herein, in light of the issues above, we propose to study a challenging yet practical problem of Geometry-aware Collective Link Prediction across Multiplex Network. To address this problem, we present a novel contrastive model, RCoCo, which collaborates intra-and inter-network behaviors in Riemannian spaces. In RCoCo, we 1 design a curvature-aware graph attention network (κ−GAT), conducting attention mechanism in Riemannian manifold whose curvature is estimated by the Ricci curvatures over the network. Thereafter, we formulate intra-and intercontrastive loss in the manifolds, in which we augment graphs by exploring the high-order structure of community and information transfer on anchor users. Finally, we conduct extensive experiments with 14 strong baselines on 8 real-world datasets, and show the effectiveness of RCoCo.

show abstract

Section: Link Predictionmentioning

confidence: 99%

“…In the literature, a series of link prediction solutions have been proposed, and concretely, there are strong methods based on random walk [1,2], subgraph patterns [3,4], graph neural networks (GNNs) [5] and the recent contrastive learning [6]. To date, most of previous works typically study link prediction in a single network.…”

Section: Introductionmentioning

confidence: 99%

Rcoco: contrastive collective link prediction across multiplex network in Riemannian space

Sun,

Li,

Yang

et al. 2024

Int. J. Mach. Learn. & Cyber.

View full text Add to dashboard Cite

show abstract

Section: Introductionmentioning

confidence: 99%

RCoCo: Contrastive Collective Link Prediction across Multiplex Network in Riemannian Space

Sun,

Li,

Yang

et al. 2023

Preprint

View full text Add to dashboard Cite

Link prediction typically studies the probability of future interconnection among nodes with the observation in a single social network. More often than not, real scenario is presented as a multiplex network with common (anchor) users active in multiple social networks. In the literature, most existing works study either the intra-link prediction in a single network or inter-link prediction among networks (a.k.a. network alignment), and consider two learning tasks are independent from each other, which is still away from the fact. On the representation space, the vast majority of existing methods are built upon the traditional Euclidean space, unaware of the inherent geometry of social networks. The third issue is on the scarce anchor users. Annotating anchor users is laborious and expensive, and thus it is impractical to work with quantities of anchor users. Herein, in light of the issues above, we propose to study a challenging yet practical problem of Geometry-aware Collective Link Prediction across Multiplex Network. To address this problem, we present a novel contrastive model, RCoCo, which collaborates intra- and inter-network behaviors in Riemannian spaces. In RCoCo, we design a curvature-aware graph attention network (κ−GAT), conducting attention mechanism in Riemannian manifold whose curvature is estimated by the Ricci curvatures over the network. Thereafter, we formulate intra- and intercontrastive loss in the manifolds, in which we augment graphs by exploring the high-order structure of community and information transfer on anchor users. Finally, we conduct extensive experiments with 14 strong baselines on 8 real-world datasets, and show the effectiveness of RCoCo.

show abstract

A Representation Learning Link Prediction Approach Using Line Graph Neural Networks

Tai,

Yang,

et al. 2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Elementary Subgraph Features for Link Prediction With Neural Networks

Cited by 8 publications

References 58 publications

Rcoco: contrastive collective link prediction across multiplex network in Riemannian space

Rcoco: contrastive collective link prediction across multiplex network in Riemannian space

RCoCo: Contrastive Collective Link Prediction across Multiplex Network in Riemannian Space

A Representation Learning Link Prediction Approach Using Line Graph Neural Networks

Contact Info

Product

Resources

About