EasyDeepLearn

GloVe vs Word2Vec.

medium

Answer

  • GloVe (Pennington 2014): explicit weighted matrix factorization of log co-occurrence counts.
  • Objective: log(Xij)\operatorname{log}(X_{\mathrm{ij}})wiw_{i}' wj  +  bi  +  bjw_{j}\; + \;b_{i}\; + \;b_{j}.
  • Batch training on global statistics (vs Word2Vec's per-window streaming).
  • Similar quality to Word2Vec; GloVe often better on analogy tasks, Word2Vec on similarity.
  • Modern: both superseded by contextual embeddings (BERT / sentence-transformers).
  • Still cited in interview questions.
Check yourself — multiple choice
  • Same thing
  • GloVe: explicit weighted log-co-occurrence factorization (global stats); Word2Vec: local windows; similar quality, both superseded by contextual embeddings
  • Random
  • Not real

GloVe: global log co-occurrence factorization; Word2Vec: local windows.

#representation-learning#nlp

Practise Unsupervised Learning

214 interview questions in this topic.

Related questions