1: \begin{abstract}

2: Graph matching finds the correspondence of nodes across two graphs and is a basic task in graph-based machine learning.

3: Numerous existing methods match every node in one graph to one node in the other graph whereas two graphs usually overlap partially in many \realworld{} applications.

4: %In this paper, a partial Gromov-Wasserstein learning framework is proposed for partially matching two graphs,

5: %which models graphs as discrete distributions and estimates the correspondence by calculating the partial tranpsort map with minimum transportation cost via a proximal point method,

6: %for which we prove convergence to a stationary point.

7: %Incorporating heterogeneous graph embedding, our framework can partially match heterogenous graphs.

8: In this paper, a partial Gromov-Wasserstein learning framework is proposed for partially matching two graphs,

9: which fuses the partial Gromov-Wasserstein distance and the partial Wasserstein distance as the objective and updates the partial transport map and the node embedding in an alternating fashion.

10: The proposed framework transports a fraction of the probability mass and matches node pairs with high relative similarities across the two graphs.

11: Incorporating an embedding learning method, heterogeneous graphs can also be matched.

12: Numerical experiments on both synthetic and \realworld{} graphs demonstrate that our framework can improve the F1 score by at least $20\%$ and often much more.

13: \end{abstract}

14: