|The Knowledge Graph matching task is to identify nodes in the two graphs that refer to the same concept. In this paper, we focus on the analysis of textual descriptions of the concepts. We employ neural language models as they can score well on text content similarity. On the other hand, we show that the text similarity of entity descriptions does not equal to referring to the exact same entity. Our text-based multi-step system was among the top participants at the Knowledge Graph matching track of the Ontology Alignment Evaluation Initiative.
*** Title, author list and abstract as seen in the Camera-Ready version of the paper that was provided to Conference Committee. Small changes that may have occurred during processing by Springer may not appear in this window.