News for hackers. Stuff that matters.

A comment on Word embeddings in 2017: Trends and future directions.

Re:Word embeddings in 2017: Trends and future directions by visarga on Saturday October 21, 2017 @09:03PM (#15524524)
No mention of StarSpace (from FaceBook) ? It figures, with the rapid pace of innovation these days.

StarSpace can compute 6 types of entity embeddings, of which word embeddings are just one type. It's a whole family of algorithms.

https://github.com/facebookresearch/Starspace/

[ Parent | on HN ]
Re:Word embeddings in 2017: Trends and future directions by tensor on Saturday October 21, 2017 @09:54PM (#15524715)
Note for those who it's relevant for, this is not usable in a commercial setting.
[ Parent | on HN ]
Re:Word embeddings in 2017: Trends and future directions by Matumio on Sunday October 22, 2017 @02:23AM (#15525439)
I don't really understand the implications of this license. Does it forbid using the resulting vectors for commercial purpose? Or does it only forbid stuff like packaging their code into a product, or offering to run it as a service.
[ Parent | on HN ]
Re:Word embeddings in 2017: Trends and future directions by anewhnaccount2 on Sunday October 22, 2017 @10:21AM (#15526939)
https://github.com/facebookresearch/StarSpace/issues/38
[ Parent | on HN ]
Re:Word embeddings in 2017: Trends and future directions by woodson on Sunday October 22, 2017 @01:08AM (#15525256)
That particular implementation isn’t (FAIR research code, non-commercial), but are there known patents on the algorithms that would prevent a clean-room implementation from being used in a commercial setting?
[ Parent | on HN ]
Re:Word embeddings in 2017: Trends and future directions by nl on Sunday October 22, 2017 @05:51AM (#15525940)
Given that this was a discussion of word vectors specifically it seems reasonable. StarSpace doesn't have anything new in terms of word vectors.
[ Parent | on HN ]
Re:Word embeddings in 2017: Trends and future directions by sebastianruder on Sunday October 22, 2017 @10:10AM (#15526897)
Good point. I've added a reference to StarSpace anyway as it's useful for applications outside of NLP.
[ Parent | on HN ]