Method and publications.
Mindojo's approach predates the current wave of AI: explicit knowledge modelling, learner models updated from every response, and content variants tested in production. Here is where it is written down.
Four commitments.
Explicit knowledge graphs
A course is authored as a graph of small units with explicit prerequisites. Because the structure is explicit, the engine can reason about it, and authors can see exactly what a learner has and has not met.
Learner modelling
Each learner has an estimated mastery for every unit, updated from every response and from how long it took. The model, not a fixed sequence, decides what comes next.
Variant testing
A unit can exist in several variants. The engine runs them against each other across sessions and keeps the ones that teach better, so content evolves with evidence.
Dialogue as the unit of instruction
Teaching happens as a conversation: a few sentences, a check, a response, a correction. This is what lets the same course run differently for every learner.
On the record.
Dynamic knowledge level adaptation of e-learning datagraph structures
US 10,373,279 B2, priority 2014, granted 6 August 2019. The method behind the engine's next-step selection over a knowledge graph.
Read the patent →MINDOJO: A platform for intelligent and adaptive online courses
INTED2014 Proceedings, Valencia, March 2014. An early description of the platform, its content model and its adaptive teaching flow.
IATED digital library →Homo Deus
“Companies such as Mindojo are developing interactive algorithms that not only teach me maths, physics and history, but also simultaneously study me and get to know exactly who I am.” Chapter 9, The Great Decoupling.
Yuval Noah Harari, Homo Deus: A Brief History of TomorrowLet's talk about your course, your school or your product.
Mindojo works with a small number of partners at a time. Tell us what you are building and we will come back with a straight answer.