I believe this to be extremely important....

[Gurtie]

hmmmm, interesting, thanks Jason!

[ergophobe]

I believe this to be extremely important....

I believe this to be extremely relevant....

The biggest advancement with RankBrain, though, is in how it deals with the quantity of content it analyzes in order to create the vectors. It seems bigger than the classic "link anchor text and surrounding text" that we always considered when discussing, for instance, how the Link Graph works.... In the patent, huge importance is attributed to context and "concepts," and the fact that RankBrain uses vectors (again, "vast amounts of written language embedded into mathematical entities"). This is likely because those vectors are needed to secure a higher probability of understanding context and detecting already-known concepts, thus resulting in a higher probability of positively matching those unknown concepts it's trying to understand in the query.

https://moz.com/blog/rankbrain-unleashed

[BoL]

Got a flashback to Teoma there.

Somewhat related, remember a good few years back G released a dump of all uni, bi, tri and quadgram words. Since then they've incorporated Freebase (a large chunk of the knowledge graph).

Theming was always a 'year away' wasn't it. Seems fairly safe to say Google has a fairly good handle on contextual relatedness (for topics there's lots of data for).