Designing automatic assessment for mathematics: trying to understand what I want as a teacher.
Chris Sangwin, University of Edinburgh, STACK Project Lead
In this talk I will discuss automatic online assessment of mathematics. This is a situation where a student answers questions and the software provides feedback, which might help the student improve or record their achievement. Typically past systems did not use artificial intelligence, instead teachers remained responsible for the design decisions. Of course, to write automatic assessments a teacher really needs to have a very clear idea of what they want to achieve, and recognise when they have achieved it! How can we randomly generate questions and still have meaningful sequences of tasks; what do these tasks hope to achieve? How can we assess a student’s answer? What is a teacher really looking for anyway? Ultimately the current solution lies in understanding what students need, in good design and in implementing that design with tool such as computer algebra.