1
$\begingroup$

I spent some time reading about both Word2Vec embeddings and alignment between different embeddings (for instance vecmap) and was wondering whether there is any significance to the word ordering of the different languages and how well the alignment can be made.

I haven't found much research in this topic (found quite a bit regarding word ordering in other cross-lingual tasks) and was wondering if the typological feature actually has an effect. Have you seen any papers in the matter?

$\endgroup$

1 Answer 1

1
$\begingroup$

Is it just about Word2Vec or NLP in general?

About Word2Vec, there is this paper that might be interesting:

Relevant Word Order Vectorization for Improved Natural Language Processing in Electronic Healthcare Records

https://arxiv.org/ftp/arxiv/papers/1812/1812.02627.pdf

Then, other ones could answer your question, but not directly. Maybe they could give some interesting clues:

Towards Structure-aware Paraphrase Identification with Phrase Alignment Using Sentence Encoders

https://arxiv.org/pdf/2210.05302.pdf

Tag-Aware Document Representation for Research Paper Recommendation

https://arxiv.org/pdf/2209.03660.pdf

Please let me know if it helps, I may have other interesting papers.

$\endgroup$
1
  • $\begingroup$ Does it answer your question? If not, please let me know. $\endgroup$ Commented Oct 29, 2022 at 14:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.