[cinntaile]

The references should allow you to follow some sort of red thread. They will probably aid in closing the knowledge gap as well. It's probably like in this article (it was a good article btw) though, you need more maths. Either you use university curricula to find out what maths you need as a starting point or you focus on the maths that pops up in the papers and work your way from there.

[tomcat27]

I can't speak for others, but I find scientific literature is not written general audience to understand, for a variety of reasons such as (a) publisher limitation on number of pages (b) to communicate core message right without distraction (c) to not bore a certain people etc.

Not to mention, depending on the sub-fields, at least in AI, there is quite a bit of noise in papers as they are explaining _their_ viewpoint which might differ from broad consensus or disproven in future.

I find papers are typically written keeping in mind a specific set of people, who are typically in that particular sub-field for some time. In sub-fields like quantum computation, that specific set of people authors have in mind while writing paper are <100, and those are not newbies.

Newbies can quickly get discouraged reading a paper out of the blue. Picking the right papers at the beginning requires exposure to the field.

It's not simple.

I really hope what you said is practical. I tried. Hope it's my fault because the alternative option doesn't exist for me.