Trjtdtk

No sign flipping while figuring out the emf of voltaic cell?

What connection does MS Office have to Netscape Navigator?

Unreliable Magic - Is it worth it?

Does falling count as part of my movement?

Plot of histogram similar to output from @risk

Why is the US ranked as #45 in Press Freedom ratings, despite its extremely permissive free speech laws?

Some questions about different axiomatic systems for neighbourhoods

Example of a Mathematician/Physicist whose Other Publications during their PhD eclipsed their PhD Thesis

Why didn't Khan get resurrected in the Genesis Explosion?

What did we know about the Kessel run before the prologues?

Why do remote US companies require working in the US?

Is there a way to save my career from absolute disaster?

Math-accent symbol over parentheses enclosing accented symbol (amsmath)

If the updated MCAS software needs two AOA sensors, doesn't that introduce a new single point of failure?

Can we say or write : "No, it'sn't"?

How to place nodes around a circle from some initial angle?

Why is quantifier elimination desirable for a given theory?

is it ok to reduce charging current for li ion 18650 battery?

I want to delete every two lines after 3rd lines in file contain very large number of lines :

Why don't programming languages automatically manage the synchronous/asynchronous problem?

Why this way of making earth uninhabitable in Interstellar?

Need help understanding a power circuit (caps and diodes)

Writing differences on a blackboard

Flying from Cape Town to England and return to another province

Understanding Youtube Recommender (item embeddings)

0

$begingroup$

https://storage.googleapis.com/pub-tools-public-publication-data/pdf/45530.pdf

section 3.2 states that

Inspired by continuous bag of words language models [14],
we learn high dimensional embeddings for each video in a
fixed vocabulary and feed these embeddings into a feedforward neural network.

1. They don't say how they created the embeddings (I guess something like https://towardsdatascience.com/using-word2vec-for-music-recommendations-bb9649ac2484)




2. I can't tell if the embeddings they created (for vidoes) is trainable in their candidate generation step or ranking step?

Embedded video watches on the above images (Are these embedding generated prior to the trainning the network?)

How about these video embeddings on the above image? (Ranking Step)

deep-learning recommender-system word-embeddings

share|improve this question

asked Mar 23 at 13:24

eugeneeugene

1064

$endgroup$

add a comment | 

0

$begingroup$

https://storage.googleapis.com/pub-tools-public-publication-data/pdf/45530.pdf

section 3.2 states that

Inspired by continuous bag of words language models [14],
we learn high dimensional embeddings for each video in a
fixed vocabulary and feed these embeddings into a feedforward neural network.

1. They don't say how they created the embeddings (I guess something like https://towardsdatascience.com/using-word2vec-for-music-recommendations-bb9649ac2484)




2. I can't tell if the embeddings they created (for vidoes) is trainable in their candidate generation step or ranking step?

Embedded video watches on the above images (Are these embedding generated prior to the trainning the network?)

How about these video embeddings on the above image? (Ranking Step)

deep-learning recommender-system word-embeddings

share|improve this question

asked Mar 23 at 13:24

eugeneeugene

1064

$endgroup$

add a comment | 

$begingroup$

https://storage.googleapis.com/pub-tools-public-publication-data/pdf/45530.pdf

section 3.2 states that

Inspired by continuous bag of words language models [14],
we learn high dimensional embeddings for each video in a
fixed vocabulary and feed these embeddings into a feedforward neural network.

1. They don't say how they created the embeddings (I guess something like https://towardsdatascience.com/using-word2vec-for-music-recommendations-bb9649ac2484)




2. I can't tell if the embeddings they created (for vidoes) is trainable in their candidate generation step or ranking step?

Embedded video watches on the above images (Are these embedding generated prior to the trainning the network?)

How about these video embeddings on the above image? (Ranking Step)

deep-learning recommender-system word-embeddings

share|improve this question

asked Mar 23 at 13:24

eugeneeugene

1064

$endgroup$

Inspired by continuous bag of words language models [14],
we learn high dimensional embeddings for each video in a
fixed vocabulary and feed these embeddings into a feedforward neural network.

share|improve this question

asked Mar 23 at 13:24

eugeneeugene

1064

share|improve this question

asked Mar 23 at 13:24

eugeneeugene

1064

asked Mar 23 at 13:24

eugeneeugene

1064

asked Mar 23 at 13:24

eugeneeugene

1064