Why do people crave a Chanel suit or a pair of Louboutin shoes? Because these iconic fashion brands create products that fit you to perfection, almost as if they were tailor-made just for you and your needs. They have redefined comfort and style.
Why Adaptive Learning?
Tailored-fit products such as a Met-Gala gown, or the less luxurious but still satisfying Spotify “made for you” playlists, are highly regarded. They offer us the satisfaction of a product that knows what you want. A way in which we translate tailored-fit products into education is through Adaptive Learning. Adaptive learning is a technique for delivering personalized learning experiences that address the unique learning preferences of individuals through a data-driven approach that self adjusts the path and pace of learning, delivering tailored fit learning experiences at scale. Learning has gone through a transformation from an established set of Learning Styles in which the learner had to fit into Learning Preferences. Learning Preferences understands that the Learner can have multiple learning styles. More than that, it understands that there is much more to learning than how content is introduced, but the interest, likes, and overall approach should not be mandated by anyone else but the learner. Not everyone fits into the same shoes. Why do we assume that everyone fits in the same learning approach?
Not everyone fits into the same shoes. Why do we assume that everyone fits in the same learning approach?
Adaptive Learning helps increase engagement and learning retention by:
- Offering learning content and challenges based on the individual level
- Ensuring that the learner can move to their nearest next level
- Avoiding the tedious repetition of training what they have already mastered
CanopyLAB’s AI-powered Adaptive Learning
CanopLAB understands that to adapt to a learner, we must understand what the learner knows, wants, and the work he or she has done. Just as you would know the size and taste in materials and colors when looking for the perfect outfit, we do the same with learning. We collect this information through:
- Learner activity: mapping the knowledge they have gained through tags, courses they have taken, their performance, and the choice they make when picking between different exercises.
- Learner’s clustering: Clustering users into groups with similar interests, activities, and performance on the platform.
- Learner performance on an adaptive entrance quiz before taking a course.
All this information is processed to create the following elements that CanopyLAB’s adaptive, social and intelligent learning space includes through AI that supports individuals with unique learning preferences:
- Designing a an excellent Adaptive Entrance Quiz
- Course builder should suggest the difficulty levels of questions
- A clear connection between questions and course materials
Designing an excellent Adaptive Entrance Quiz – What starts well ends well
One major problem to deal with when building an adaptive system is facing a “cold-start”. If a learner is new to the platform or a course is just created, there is no prior information to base recommendations on. This is why we have included an Adaptive Entrance Quiz. This is a way of identifying from the start of the course information regarding new learners and gathering some more information about “old” learners on new topics. You want to get to know your learners as much as possible at the beginning; a good and informed start is key to a well-executed course.
If a learner is new to the platform or a course is just created, there is no prior information to base recommendations on. This is why we have included an Adaptive Entrance Quiz.
The purpose of adaptive entrance quizzes is to detect if a user has already mastered some of the course’s topics. Questions are grouped based on topics, ranked by difficulty. If a learner fails to answer a few questions from one topic, the AI will assume that the learner does not know about that topic yet. Therefore it will skip that and move on to the next topic. By doing this, we prevent forcing learners to work on something they are not familiar with yet at the very beginning of the learning process.
For course creators, these are guidelines that will assure that you conduct an effective entrance quiz that concludes invaluable and applicable information:
- Entrance quizzes have to cover all topics in a course to understand the level of learners’ knowledge regarding the topics to be included
- For each topic, a sufficient number of questions is required to conclude what is the level of each learner in each topic
- Questions should be clear and focused. Important questions should be highlighted
The course builder should suggest the difficulty levels of questions – David and Goliath is an old tale, not a reality
David and Goliath is an excellent example of how someone “small” beats a huge challenge. However, this is an exception and should not be a rule of thumb. Learners should be faced with questions that accommodate their skills and capacities. Let’s put Goliaths with Goliath difficult questions while we help the Davids become Goliaths with some David-level questions. This is done in CanopyLAB’s learning space through Difficulty Format Adaptive Quizzes.
These are the general guidelines to create useful Difficulty Format Adaptive Quizzes:
- Without prior information about a new user or a new course, the course creator should suggest a difficulty level
- When a learner takes a quiz, questions should be shown from easy to difficult. Offering a very difficult question at the beginning might make them frustrated and take away any motivation left to complete it
Although we ask course builders to suggest the initial difficulty level based on their perspective, the difficulty level of questions to each learner might not be the same, currently, the AI system can automatically suggest difficulty levels for questions in an adaptive quiz through two different ways:
- Ranking difficulty based on learner’s prior knowledge: The system will review the knowledge already acquired, its performance on other courses and exercises, and will group learners that have the same knowledge, performance and use this cluster as a base for performance on this particular quiz.
- Ranking based on the performance of learners who have completed a quiz: Difficulty level can be calculated after several learners have completed the quiz. The suggested level of difficulty reflects the average performance of learners on that quiz. This method is helpful for when a new learner takes a quiz, as we do not have as much information as for that, we can use peers’ performance to rank difficulty.
The result is a quiz that adapts to the difficulty of the questions, avoiding frustration and creating a progress system accommodated to the learner’s needs.
A clear connection between questions and course materials – Mixing apples with oranges is never a good idea; why does it happen in a quiz?
I’m not alone in expressing my frustration when, during a quiz, I was asked about something that I was never taught. Sometimes it was the teacher’s fault, some others, it was mine for not paying attention to all of my course materials. Regardless of whose fault it is, the morale is that course materials should be easily accessible for learners, and the questions selected should reflect what has been made accessible to users.
AI can help to form a clearer link between the materials and the courses in different ways:
- Automatic creation of exercises. Let AI create exercises based on your course materials, make sure that comprehension is tested correctly and in an efficient manner.
- An ever-present support system. When learners get stuck in a question/level, the system can suggest reviewing some knowledge from a course based on the links between questions and materials.
Let AI create exercises that adapt to what you are teaching, and make sure you test them on what you are teaching. Make sure that they comprehend your course materials and feel empowered by offering questions that they can answer and to which they already have the resources where to find an answer.
T(AI)lored-fit learning available for everyone
The moral of the story is do not offer your learners a one-size-fits-all approach to learning. 21st-century education promises custom-made learning like a perfect fit Chanel suit. Adaptive learning and building adaptive learning experiences are the least they can expect. Drive learner engagement and increase learning retention by creating personalized learning experiences. Let’s be honest if you could, would you rather choose a standard size suit, or one made just for you?
So what are you waiting for? You already have some pro tips and insights on how we do it, do not wait to offer your learner’s an adaptive learning experience. We want to help you help your learners. Do you have your own CanopyLAB Learning space? Yes? Go ahead and try this out. No? Let us help you help your learners, take the Tesla of education out for a free test ride, book a demo.
