[hiker512]

Can you recommend an implementation or paper for handling OOV words via character level embeddings?

[patelajay285]

We just open sourced a easy to use library called Magnitude that handles out-of-vocabulary words and uses Annoy indexing for fast most_similar queries for word2vec, GloVE, and fastText:

https://github.com/plasticityai/magnitude

[Sinidir]

Thank you very much. This might help with my Master Thesis :)

[visarga]

Great work!

[fnl]

I agree - and also to your initial comments on the actual benefits & limitations of word vectors, that I can 100% subscribe to.

[wenc]

Just an FYI, FastText's default implementation handles OOV words via word n-grams and character n-grams. (see switches -minn, -maxn, and -wordNgram)

https://fasttext.cc/docs/en/options.html