News for hackers. Stuff that matters.

A comment on Show HN: Sense2vec model trained on all 2015 Reddit comments. Parent comment

Re:Show HN: Sense2vec model trained on all 2015 Reddit comments by syllogism on Monday February 15, 2016 @06:08PM (#11106760)
?

https://sense2vec.spacy.io/?Barack_Obama%7CPERSON

https://sense2vec.spacy.io/?Obama%7CPERSON

[ Parent | on HN ]
Re:Show HN: Sense2vec model trained on all 2015 Reddit comments by gleb on Monday February 15, 2016 @06:11PM (#11106772)
I get: https://sense2vec.spacy.io/?Obama%7CORG

First result is GOP

[ Parent | on HN ]
Re:Show HN: Sense2vec model trained on all 2015 Reddit comments by nl on Monday February 15, 2016 @06:40PM (#11106919)
It is showing things which are used in English in the same way as "Obama" when the word "Obama" is used as an organization.

Clearly "Obama" isn't an organization, but it does a decent job. Consider the sentence "Obama’s unsolicited advice that Congress should “go ahead and vote” has only hardened the GOP majority’s resistance."

[ Parent | on HN ]
Re:Show HN: Sense2vec model trained on all 2015 Reddit comments by henningpeters on Monday February 15, 2016 @06:15PM (#11106797)
Try to search for Obama|PERSON
[ Parent | on HN ]