Structural representation learning for network alignment with self-supervised anchor links
作者：
Highlights：
• Our approach is scalable to real-world online social networks.
• Our method is robust to structural noises up to 40%.
• Our model outperforms unsupervised methods by over 13% accuracy.
• Our model is cheaper than supervised methods in not using any prior knowledge.
• Our performance is robust and maintains 90% accuracy even if data is 10% sparser.
摘要
•Our approach is scalable to real-world online social networks.•Our method is robust to structural noises up to 40%.•Our model outperforms unsupervised methods by over 13% accuracy.•Our model is cheaper than supervised methods in not using any prior knowledge.•Our performance is robust and maintains 90% accuracy even if data is 10% sparser.
论文关键词：Graph mining,Graph matching,Network alignment,Network representation learning,Network embedding
论文评审过程：Received 28 June 2019, Revised 22 July 2020, Accepted 6 August 2020, Available online 18 August 2020, Version of Record 7 September 2020.
论文官网地址：https://doi.org/10.1016/j.eswa.2020.113857
