Cygnet: Refactoring the Open Multilingual Wordnet

Abstract

The wordnet file specification empowers the creators of different wordnets by allowing them to encode the same information in multiple different ways. A drawback of this approach is that redundancy is introduced. As a consequence, different wordnets often contain conflicting records, creating issues when one attempts to conduct multilingual research using multiple wordnets simultaneously. To address this, we present the OMW Cygnet, an experimental reformulation of wordnet that is designed to eliminate conflicting records and improve modularity. We convert data in 47 languages from the Open Multilingual Wordnet into this format, and release a web browser which makes it easy to navigate multilingual wordnets.

Resources

Details