FANDOM


Retrofitting: take word vectors from a distributional semantics model and fit them to WordNet hierarchy (synonym, hyper-/hyponym) while not changing them too much.

Fried & Duh (2015)[1]: incorporate WordNet into training, cosine distance is aligned with path-based distance.

References Edit

  1. Daniel Fried, Kevin Duh. 2015. INCORPORATING BOTH DISTRIBUTIONAL AND RELATIONAL SEMANTICS IN WORD REPRESENTATIONS. PDF