Arbital path

https://arbital.com/p/Arbital_path

by Alexei Andreev Jan 27 2016 updated May 11 2016

Arbital path is a linear sequence of pages tailored specifically to teach a given concept to a user.


When a user wants to learn a particular concept, Arbital can use the existing "requires" and "teaches" relationships to generate a list of pages for the user to read. This list will present pages in order, so that the reader can build on the knowledge they already have, and by the end of it the reader will have all the necessary requirements to understand the concept they are trying to learn.

The algorithm

It will be easier to create paths if you understand Arbital's underlying algorithm. Here are the key steps:

  1. Gather the list of the requisites the user wants to learn.
  2. Remove the ones the user already knows.
  3. For each requirement, find a page that teaches it and add it to the path. (There is a somewhat complex scoring function for picking the best teaching page.)
  4. For each of those pages, take their requirements and repeat from step #2.
  5. Once the entire tree is generated, Arbital linearizes it, so that every requisite is taught before pages that require it. (If there are dependency cycles, they are broken arbitrarily.) Note that as long as the requirement constraints are satisfied, the pages can otherwise appear in any order.

On the /learn/ page the resulting pages are displayed in a sequence. If you want to see the underlying tree structure, use the ⋮ menu to switch the view. It should make it more clear why the algorithm chose certain pages in certain order, and you can also explore the alternatives.

For a demo, take a look at: Bayes' rule: Guide.

Debugging

Here are a few likely issues you might run into and tips on how to address them.

The order of the pages is not correct:

The algorithm is picking another page instead of yours:

If all else fails, report your issue to us, and we'll help you!