Can someone direct me to research papers that have tried to understand word embeddings?
The original papers written by Thomas Mikolov are word2vec and doc2vec.
Lecture notes given by Richard Socher is simple to understand.
word2vec explained gives an intuition how the negative sampling in word embedding works. Not the whole algorithm though.
Nevertheless, there are implementations in various languages and libraries explaining how it works in their docs.
Python - gensim, Tensorflow
java - deeplearning4j
$\begingroup$ thanks but after the original papers...do you know if there were any other research papers that tried it. $\endgroup$– devcJun 1, 2017 at 16:21
1$\begingroup$ @devc: Glove tries to generate word embeddings too, but in a different approach. nlp.stanford.edu/pubs/glove.pdf $\endgroup$– chmodsssJun 1, 2017 at 21:46
$\begingroup$ See also research.fb.com/fasttext $\endgroup$– EmreAug 30, 2017 at 16:15