The platform works by extracting a list of subjects that correspond with topics in the learning materials and then organizing the subjects by size into different layers. The subjects are linked to the material, and some materials can be linked to more than one topic.
For example, the top layer of the process could be science, technology, engineering and math (STEM). The next layer could be divided into physics and biology.
If a calculus student needs to sort through the thousands of open educational resources related to STEM just for materials about derivatives, a guided pathway is then created through the STEM layer, into the physics and biology layer, where it ignores the biology path and leads the student onto the next layer.
At this layer, the student can find differential calculus lecture notes and physics videos focusing on mechanics.
By using detailed learning behavior technology created by MIT and Fujitsu, educators and learning system providers can predict learning outcomes through simulations without having to rely on assessing the thousands of students who are enrolled in a large online course.
These algorithms can help system providers determine the best suited material for such a large amount of students, said Fujitsu’s director of business development Surya Kumar Josyula, further helping learners and teachers find the right material faster.
“This engine simulate environments, tests out what’s working and what’s not, without even the having students there,” Josyula said. “But at the same time the guided pathways are personalized down to the individualized student level, which typical learning engines do not do.”