Adaptive learning and Open education: A Match made in heaven?

What is adaptive learning? When reading on the subject many forms of definitions appear that also determine most of the pro’s and con’s. Below you see three examples of definitions from different points of view: 

• Adaptive learning is an educational technology that uses algorithms to adjust the presentation of material to an individual student's needs, based on their performance and other data.
• Adaptive learning is a method of instruction that personalize the learning experience for each student, by adjusting the content, pace, and difficulty level of the material to meet their individual needs.
• Adaptive learning is a process by which learners actively participate in the design and delivery of their own learning experience, by providing feedback to the system and adjusting their own learning goals and strategies.

We believe understanding and looking at adaptive learning from the various viewpoints is key for its success and will keep the right balance between what the technology does, together with the lecturer and learner. Our vision of adaptive learning is focusing on increasing the accessibility of learning while also increasing the impact of understanding knowledge to make it easier for teachers and students to make conscious decisions based on insights. Adaptive learning should not just be a tool or AI telling you what to do. Yet in complex learning, like with maths and statistics, a tool giving you insights based on a combination of massive data is helpful.

At Grasple we based our adaptive learning on the following design principles:

1. It should enhance open education and increase accessibility 
2. It is a tool to help learners in their personal learning path
3. It should provide a high level of autonomy: where teacher and student are in charge of the content and choices which helps them staying motivated and grasping the subject matter as a crucial part to stay motivated and grasp the subject matter 

We believe the combination of open educational resources (i.e. open mathematics/statistics exercises) and adaptive learning has a high potential for creating a learning impact for students. We strive to make learning mathematics and statistics more accessible for everyone around the world (improving equity). To do so we already make sure the content is openly licensed so that everyone can always access/use/copy/improve it and not one party is in control of the (copyrighted) content. For the future we also want to make sure you can interact with the content via an adaptive learning method as an individual without having to pay for that (or the access of the content).

To enhance the potential of adaptive learning and open education we will focus on the next level of cooperation within the community and enhancing content quality by creating a method for teachers to be able to improve, update, maintain and extend knowledge component graphs (KCG) within the community so that openly licensed exercises can easily be used in the adaptive method.

This approach will tackle some of the challenges mostly associated to the development of adaptive learning like:

• The requirement for high volume of quality content
• The low level of autonomy for teachers and students
• Absence of transparency and insights in the algorithm (i.e. black box)
• Increase in content learning material costs for students 

How does Grasple apply adaptive learning in practice and tackle some of the difficulties and concerns?

When focusing on technology major concerns are often: Who is in control? How does the algorithm make decisions? Should the student fail or pass? At Grasple we believe that the technology's power lies in collecting data and combining it with expert knowledge from the teacher and student to provide insights and advice while still letting the student/teacher make the final decision.

Grasple developed three key features in its platform:

• Knowledge component graphs (KCGs)
• Diagnostic testing 
• Easy content creation, sharing and editing moving towards interoperability

The knowledge component graph (KCG) is developed as a type of graphical representation used to model the relationships between different knowledge components in a given domain. Knowledge components are the basic building blocks of knowledge that a learner needs to acquire in order to become proficient in a particular subject area.

Grasple developed the graphs with teachers and learning experts. We can create KCGs for a course and use them in tests manually. In the future we strive for teachers to be able to do this themselves based on a base KCG. The learnings and insights from these can be openly shared just like Grasple facilitates with open learning resources. The KCG is not just meant to force a learner into a path. It provides insight and advice on what next learning steps can be taken, to both the teacher and students. Teachers have a say in the design of the KCGs for the subjects and create the model with their expertise of the subject, learning audience and learning goals for their courses. The system adds data and enhances the expertise of the lecturers. 

The subjects and learning goals connect in the graph via two relations: Hierarchy or prior knowledge. The hierarchical relationships show how a bigger topic is split into its subtopics and indicate that someone's knowledge about the subtopics will feed into their knowledge about the overarching topic. Prior knowledge relations indicate a certain knowledge needed before someone can continue with the next subject.

The relationships between the knowledge components can estimate the mastery of a subject while simultaneously revealing gaps in knowledge.

A diagnostic test is developed for courses based on subjects in the course. Again these can be shared and leveraged within the community of users and even exported as Grasple is an open platform. The diagnostic test provides insights to the student and teacher on where we think you can use more practising. The information is updated based on the exercises a student completes. The exercises are curated by the lecturers responsible for the course (teacher autonomy).

The student in the course is not forced to learn within a fixed path set by an algorithm. The system collects information and creates insights. The student can still choose where to spend their time (student autonomy). 

Working with the community content saves time in content creation. In the end the benefit for the teacher is seeing your students learn and grow based on insights and being able to make conscious decisions on how to structure their offline lecture, and how to balance online with offline activities where all students are able to participate and gain learning value.

N.B. Detailed insights only available in the institution account not for individual accounts due to high privacy and security demands

A challenge for adaptive learning is having a lot of data and fine meshed content available. With the creation of an open resources platform the Grasple community creates a lot of content with fine meshed feedback embedded in the exercise. The statement “many hands make light work” counts in Open Education. Teacher’s workload is high and the effort they put in creating content is not always seen in their appraisals. Being able to make use of a community and cooperate within a user-friendly editor enhanced the adoption of open resources within institutions and by individual teachers. Individual teachers can use Grasple for free when creating content for the community while teaching learners. 

Combining the above with an easy to use platform for teachers will drive for impacting a large group of learners in the world since they will have free access to openly licensed interactive math/statistics exercises maintained and improved by a large group of teachers from around the world and use those materials in an interactive adaptive way such that it facilitates them in learning at their own pace and level of mastery.

At Grasple we continue to work on the following; 

• Use expert driven KC graphs together with AI research on student answer data to continuously determine which open exercises are best to use in the adaptive learning method (the research part) 
• Allow teachers to easily create/adapt KC Graphs for their courses/materials
• Make the KC Graphs part of the open content in terms of collaboration and open licensing
• Have community KC graphs (like we have community materials now) 
• Have an adaptive way of interacting with the open exercises using those KC graphs and the open exercises for individual users (instead of the linear way of interacting with them now)
• Create a community for sharing open pedagogic strategies, a.o. balancing on- and offline learning. 

Please send us your insights and/or comments on adaptive learning. We love to hear more ideas and concepts to sharpen ours and other visions on making knowledge available for all.