In recent years, powered by the learned discriminative representation via graph neural network (GNN) models, deep matching methods have made great progresses in task of semantic features. However, these usually rely on heuristically generated patterns, which may introduce unreliable relationships to hurt performance. this paper, we propose a joint learning and network, named GLAM, explore relia...

Evaluation of the closeness of two texts is a subtask for F T R and I R systems. The basic means used to accomplish it is the matching of a to m ic te x t e n t i t ie s (ATEs) such as words, stems, simple phrases and/or concepts. We address the question how concepts can be used as ATEs more efficiently in order to match “s m a l l d u c k ” with “s m a l l b ir d ”. The o n to -m a tc h in g t...

The most common matching applications, e.g., ontology matching, focus on the computation of the correspondences holding between the nodes of graph structures (e.g., concepts in two ontologies). However there are applications, such as matching of web service descriptions, where matching may need to compute the correspondences holding between the full graph structures and to preserve certain stru...