Deep Recursive Network Embedding with Regular Equivalence

Network embedding aims to preserve vertex similarity in an embedding space. Existing approaches usually define the similarity by direct links or common neighborhoods between nodes, i.e. structural