There is an old, tried-and-tested learning algorithm called SR (Spaced Repetition)
(The algorithm used to determine when to show cards in flashcard programs)
The problem is that SR is a tool that only fits 1:1 structures,
and this structure does not match the structure of knowledge
What fits this structure is as simple as country–capital pairs or word–meaning pairs,
but not general subjects
Now, LearnObit has a solution for that!
The first step is to make a good note-taking tool
that mimics the tree structure for the purpose of organizing knowledge
The second step is to connect that note-taking tool with SR
Then SR can be easily used for learning information that has tree structures as well as 1:1 pairs
And since the tree structure is the structure of general subjects,
SR can be used for general subjects unlike in the past
Now SR is the normal way, not just for a few enthusiasts on the fringes
One thing stood out to me in your explanation — you say “tree structure” is a natural way to organize knowledge.
Why is that the case? I generally use one of the modern “tools of thought” (first Roam and now mem.ai) and in these tools, you just dump notes and add tags and and links among notes, so you have a general graph structure (from the links) overlayed with tags.