Efficient Estimation of Word Representations in Vector Space

We propose two novel model architectures for computing continuous vector representations of words from very large data sets. The quality of these representations is measured in a word similarity...