Re:Collaborative filtering doesn't work for us
by caramelazimuth on Friday August 20, 2021 @07:58PM
(#28252730)
How does using these embeddings (of sequences of interactional partners) compare with using embeddings of words/tokens?
[ Parent
| on HN ]
Re:Collaborative filtering doesn't work for us
by kevin948 on Friday August 20, 2021 @09:12PM
(#28253174)
They should be quite similar.
In the end you coax your embedding space to amount to some consistent measurement of what causes samples to diverge from one-another.
You can do similarity search and all the sorts of things you do for word embeddings on embeddings generated for other scenarios.