Welcome to Amplifire

Adaptivity Intellectual Property Explained

An adaptive system is one that focuses learners’ time on what they need and provides individualized assistance when necessary. Adaptivity is the solution to ineffective, “one-size-fits-all” learning. For instance, imagine a room full of learners watching a video; some are bored, some are confused, and very few are getting exactly what they need.

Why do it? What’s the value?

• Learners spend time only where they need to.
• Learners with lots of prior knowledge, or who learn quickly, can complete modules in a small fraction of the time taken by the average learner. An adaptive system can prevent the boredom experienced by learners who are ahead of the curve.
• Learners who struggle or need extra guidance receive it. An adaptive system prevents the frustration experienced by learners who couldn’t otherwise keep up.
• If the client uses a blended-learning or instructor-led approach, they can accommodate more learners per instructor. Adaptive systems shoulder some of the work, letting the instructor efficiently handle more trainees/students. This can reduce costs by minimizing the burden on instructors; with the same enrollment, fewer instructors are needed.

What differentiates how Amplifire Works?

• One way to adapt is to use a simple item-response approach to adaptivity. This means that if you get a question wrong, the system asks you an easier one; if you get a question right, the system asks you a harder one.
• Another approach to adaptivity is to use more advanced algorithms (e.g., Bayesian models) to estimate what learners know versus don’t, and what they should see sooner, later, or never. Bayesian systems ask learners questions to determine the state of their knowledge and direct them to more basic or more advanced content to fill in knowledge gaps.
• In Amplifire, every learner encounters every item. Especially in the context of the other approaches, that probably doesn’t seem very adaptive! Instead, Amplifire adapts in several other ways.
  1. How long each learner needs to spend on each item. When a learner already knows something, they only spend as long as it takes to answer the relevant question—often only a few seconds. Amplifire devotes more time to content where learners are struggling or where their metacognition is skewed.
     • Added benefit: Because Amplifire asks every person every question, our reports can provide diagnostic information about what the entire population knew about each concept before training began.
  2. What a learner needs to see next. Amplifire decides what item a learner should next encounter. This algorithm shuffles content in ways that best benefit learners.
  3. Whether to provide corrective feedback. Everyone sees “valence feedback,” which tells a learner whether they were right or wrong. This gamification boosts learners’ dopamine, causing better retention and more engagement. 
  4. How long to wait before providing corrective feedback. If the learner wasn’t confident and correct, Amplifire does not immediately correct them. 
  5. Whether and how long the platform waits to ask a question again. If a learner has become confident and correct and their prior responses do not indicate that the item will quickly slip out of memory, the system will consider that item mastered. If the learner needs to see the item again, Amplifire harnesses the spacing effect (Landauer & Bjork, 1978) by waiting to repeat the item.
  6. Whether to issue meta-guidance to the learner. A learner who is rushing through can receive a message to slow down; someone who is starting to drag and struggle can receive a suggestion to take a break. A learner who has fallen behind on their assignments can receive an email prompting them to catch up.