I use Princeton's WordNet in NLP applications. I always read that it is "human-curated." But how is it curated? Who decides that words belong in a synset? How does the process work? Wikipedia does not give many details. How does it work?
As far as I know, they do the same way they created things like Penn tree bank. They have a panel of experts, usually linguists, who sat together and decided synset taxonomy based on various widely used dictionaries.
PS: Please up-vote me so I can up vote others, thanks :)