BSAL: A Framework of Bi-component Structure and Attribute Learning for Link Prediction

Abstract

Given the ubiquitous existence of graph-structured data, learning the representations of nodes for the downstream tasks ranging from node classification, link prediction to graph classification is of crucial importance. Regarding missing link inference of diverse networks, we revisit the link prediction techniques and identify the importance of both the structural and attribute information. However, the available techniques either heavily count on the network topology which is spurious in practice, or cannot integrate graph topology and features properly. To bridge the gap, we propose a bicomponent structural and attribute learning framework (BSAL) that is designed to adaptively leverage information from topology and feature spaces. Specifically, BSAL constructs a semantic topology via the node attributes and then gets the embeddings regarding the semantic view, which provides a flexible and easy-to-implement solution to adaptively incorporate the information carried by the node attributes. Then the semantic embedding together with topology embedding are fused together using attention mechanism for the final prediction. Extensive experiments show the superior performance of our proposal and it significantly outperforms baselines on diverse research benchmarks.

An interpretable ensemble method for deep representation learning

“…The field of neural networks has seen a growing interest in the incorporation of hyperbolic geometry (Peng et al, 2021;Yang et al, 2022b;Zhou et al, 2022b;Vyas et al, 2022;Xiong et al, 2022a), in areas like lexical entailment (Nickel & Kiela, 2017;Gulcehre et al, 2019;Sala et al, 2018), knowledge graphs (Chami et al, 2020;Bai et al, 2021;Sun et al, 2020;Xiong et al, 2022b), image understanding (Khrulkov et al, 2020;Zhang et al, 2020;Atigh et al, 2022;Hsu et al, 2021), and recommender systems (Vinh Tran et al, 2020;Chen et al, 2022;Sun et al, 2021a;Yang et al, 2022c;a). In the realm of graph learning (Gulcehre et al, 2019;Chami et al, 2019;Liu et al, 2019;Yang et al, 2022b), a significant amount of research works generalizing graph convolutions (Kipf & Welling, 2017;Veličković et al, 2018;Hamilton et al, 2017;Yang et al, 2020;2022d;Li et al, 2022;Zhang et al, 2019) in hyperbolic space for a better graph or temporal graph representation (Chami et al, 2019;Liu et al, 2019;Zhang et al, 2021b;Yang et al, 2021b;Bai et al, 2023;Sun et al, 2021b), which has achieved impressive performance.…”