WL-Align: Weisfeiler-Lehman Relabeling for Aligning Users across Networks via Regularized Representation Learning

Keyphrases

• Alternating Phases 20%
• Anchor-based 20%
• Delabeling 100%
• Embedding Space 20%
• Exact Matching 20%
• Graph Representation Learning 20%
• Hashing 40%
• Individual Networks 20%
• Label Propagation 20%
• Learning Framework 20%
• Learning Module 20%
• Long-range Connectivity 20%
• Low-dimensional Embedding 20%
• Node Representation 20%
• Performance Improvement 20%
• Precise Alignment 20%
• Regularized Representation Learning 100%
• Representation Learning 40%
• Second-order Proximity 20%
• Similarity-based 20%
• State-of-the-art Techniques 20%
• Synthetic Data 20%
• Weisfeiler-Lehman 100%

Computer Science

• Embedding Space 25%
• Graph Representation Learning 25%
• Hashing 25%
• Learning Framework 25%
• Performance Improvement 25%
• Representation Learning 100%
• Synthetic Datasets 25%