Multi-level Graph Convolutional Networks for Cross-platform Anchor Link Prediction

Abstract

Cross-platform account matching plays a significant role in social network analytics, and is beneficial for a wide range of applications. However, existing methods either heavily rely on high-quality user generated content (including user profiles) or suffer from data insufficiency problem if only focusing on network topology, which brings researchers into an insoluble dilemma of model selection. In this paper, to address this problem, we propose a novel framework that considers multi-level graph convolutions on both local network structure and hypergraph structure in a unified manner. The proposed method overcomes data insufficiency problem of existing work and does not necessarily rely on user demographic information. Moreover, to adapt the proposed method to be capable of handling large-scale social networks, we propose a two-phase space reconciliation mechanism to align the embedding spaces in both network partitioning based parallel training and account matching across different social networks. Extensive experiments have been conducted on two large-scale real-life social networks. The experimental results demonstrate that the proposed method outperforms the state-of-the-art models with a big margin. CCS CONCEPTS• Information systems → Data mining.

Contrastive Multi-View Multiplex Network Embedding with Applications to Robust Network Alignment

“…Extensive experimental results on real-world benchmarks show the state-of-the-art performance of our approach. LINE 6 [37] models first-order and second-order proximity, trains them separately with edge sampling, and concatenates the representations after normalization on node embedding. The number of samples is set as 3 × 10 8 .…”

“…The inputs are the same as FINAL. 6 https://github.com/tangjianpku/LINE 7 https://github.com/THUDM/ProNE 8 https://github.com/ColaLL/IONE 9 https://github.com/maffia92/FINAL-network-alignment-KDD16, for both FINAL and IsoRank…”

“…Network alignment [6,53] aims to predict anchor links between layers of multiplex networks. As a common protocol in line with [8,21,45], we randomly choose a specific portion of anchor links as the known seed matchings across layers, and predict the rest of anchor links according to the learned embeddings {X (д) } N д=1 .…”

Multi-level Graph Attention Network based Unsupervised Network Alignment

“…DALAUP [7] uses a pair of neural networks with shared parameters to obtain the vector representation of user pairs, and integrates three query strategies to select the user pairs with the most abundant information for labeling and model training. MGCN [6] takes both local and hypergraph level graph convolutions into consideration to learn network embeddings. And it develops a series of treatments including network partitioning and space reconciliation to handle the distributed training process.…”

“…GCN [11] simplifies the previous method by restricting the filters to operate in a 1-step neighborhood around each node. MGCN [6] uses multi-level graph convolutional networks to capture wider information for network alignment. However, the weights assigned to different neighbors by the above method in the neighborhood of the same order are exactly the same, which limits the ability to capture the correlation of spatial information.…”