Relation to Glove and word2vec

[alexanderpanchenko]

Add a description (in one paragraph) which relates our model (in terms of the math notation already introduced in the paper) to the Glove https://nlp.stanford.edu/pubs/glove.pdf and word2vec.

Glove:

[alexanderpanchenko]

rewrite Glove and SkipGram objectives here in our notation + add a few sentences about this.

[alexanderpanchenko]

[alexanderpanchenko]

https://arxiv.org/pdf/1402.3722.pdf

[alexanderpanchenko]

@m-dorgham any updates on this?

[m-dorgham]

@alexanderpanchenko Have a look at this paragraph: https://www.overleaf.com/9629192926rfsnkrbwyyyt

Is this enough or do I need to explain more?

[alexanderpanchenko]

In the last equation instead of exp should be sigma??

Sent from my iPhone

On 8. Dec 2018, at 15:13, Mohammad Dorgham notifications@github.com wrote:

@alexanderpanchenko Have a look at this paragraph: https://www.overleaf.com/9629192926rfsnkrbwyyyt

Is this enough or do I need to explain more?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[m-dorgham]

The equation with the sigma is the final objective function after introducing the concept of negative sampling. The one I wrote is the general objective function (this is how it was written in the GloVe paper as well). We can write it with the sigma notation or leave it in the general form, as you like. But if we wrote it in the sigma notation we might need to explain in more words the concept of negative sampling.

[alexanderpanchenko]

i think we need to write the most concrete one with the negative samplings as we do the same

On Dec 8, 2018, at 5:37 PM, Mohammad Dorgham notifications@github.com wrote:

The equation with the sigma is the final objective function after introducing the concept of negative sampling. The one I wrote is the general objective function (this is how it was written in the GloVe paper as well).