4 ms·
That's what word2vec is trying to model, it cannot do it because natural language semantics are person-2-world relationships. It groups words together based on
by mjburgess 9y ago
That's what word2vec is trying to model, it cannot do it because natural language semantics are person-2-world relationships.
It groups words together based on their distributional characteristics, and with the right parameter tuning, that can be a good-enough model of similarity.