It is the night before the exam and you have four hours. Which chapter do you open? Most students answer that question with whichever topic feels scariest — which is not the same as the topic they are actually weakest at. The knowledge map exists to answer it properly.
What gets built
Alongside your course, an agent extracts the concepts in it and how they depend on each other into a map. Each node carries a short summary and a pointer back to where in the course it is taught — so a weak spot on the map is one click away from the paragraph that fixes it. No hunting through the chapter for the bit you didn't get.
Four scores, not one percentage
A single 'mastery' number hides the thing you most need to know: which kind of not-knowing you have. So every concept is scored on four separate axes.
- Memory — can you recall it at all?
- Comprehension — do you actually understand what it means?
- Structure — do you know how it connects to the rest of the chapter?
- Application — can you use it on a question you have not seen before?
Memorised but can't apply it is a completely different problem from understood but keeps forgetting. One percentage cannot tell you which one you have — and they need opposite fixes.
Where the scores come from
They are not self-reported, and they are not just your MCQ percentage. After each study session — an MCQ set, a flashcard run, or a Live Tutor attempt where you explain the idea back in your own words — an evaluator agent reads what you actually did and updates the affected concepts, with a note explaining why it moved them.
It does not lock anything. There is no checkpoint that blocks the next part until a score clears a bar, and no schedule nagging you to come back. The map tells you where you are weak; what you do about it is entirely yours.