Replace digits with # in any term where a sequence of two digits appears. This operation is applied to text that passes through word2vec, so

Tokenize text using the default wordfreq rules.

Given a list of tokens, remove a small list of English stopwords. This helps to work with previous versions of ConceptNet, which often provid

Get a URI that is suitable to label a row of a vector space, by making sure that both ConceptNet's and word2vec's normalizations are applied