Papers with DeepAlignment
DeepAlignment: Unsupervised Ontology Matching with Refined Word Vectors (N18-1)
Copied to clipboard
| Challenge: | Ontologies compartmentalize types and relations in a domain and require a process to establish alignments between entities to unify and extend existing knowledge. |
| Approach: | They propose a method which refines pre-trained word vectors to derivate ontological entity descriptions tailored to the ontology matching task. |
| Outcome: | The proposed method improves ontology matching performance over the current state-of-the-art. |