An adaptive learning platform is software that picks each child's next question from that child's last answers. A child who struggles gets easier practice and a short explanation, and a child who is ahead moves on to harder work. The method is explained in adaptive learning. Schools buy these platforms for maths, language and science practice, in class and as homework.
A fixed worksheet gives every child the same 20 questions. For a child who has not understood the topic, most of the questions are too hard, and the child learns little. For a child who already knows the topic, they are too easy. An adaptive platform tries to keep each child working at the edge of what that child can do.
How does it work? The platform has a large bank of questions. Each question is tagged with the skill that it tests and with how hard it is. The platform keeps an estimate of how well each child knows each skill. After every answer, it updates the estimate and picks the next question. A good platform also studies which wrong answer the child chose, because a wrong answer often shows a particular mistaken idea.
Imagine a Class 5 student in Guwahati practising decimals for 20 minutes. The platform asks which is larger, 0.5 or 0.25. He chooses 0.25. This is a common mistake: the child thinks that 25 is larger than 5, so 0.25 must be larger than 0.5. The platform does not simply mark him wrong and move on. It shows him a short explanation with a strip divided into ten parts and a strip divided into a hundred parts. Then it gives him three easier questions on the same idea. He gets all three right. So it returns to the original kind of question, and he gets that right too. A classmate who answered the first six questions correctly has already moved on to adding decimals.
The teacher sees the result for the whole class the next morning. The report shows, for example, that 14 of 38 children have the same mistaken idea about decimals. So she teaches it again to the whole class instead of to each child separately. Where to begin is often decided by a diagnostic assessment, a short test that shows what each child has not yet learned. Each child's progress can also go into the school's student information system (SIS), which holds attendance and marks.
The topics that the platform practises usually follow the chapters that the teacher has just taught from the digital courseware.
How is an adaptive platform different from an AI tutor? An adaptive platform chooses from a fixed bank of questions that teachers and subject experts have written and checked. An AI tutor holds a conversation and answers any question that the child types, in the child's own words. Newer products combine the two: the adaptive platform chooses the practice, and an AI tutor explains a step when the child asks.