AI RESEARCH
Towards Metric-Faithful Neural Graph Matching
arXiv CS.LG
•
ArXi:2605.06588v1 Announce Type: new Graph Edit Distance (GED) is a fundamental, albeit NP-hard, metric for structural graph similarity. Recent neural graph matching architectures approximate GED by first encoding graphs with a Graph Neural Network (GNN) and then applying either a graph-level regression head or a matching-based alignment module. Despite substantial architectural progress, the role of encoder geometry in neural GED estimation remains poorly understood.