Colleagues have developed a new web platform to deliver automated feedback on student homework. Led by Dr Peter Johnson, the platform – called Lambda Feedback – is currently in its ‘Alpha version’. Thus far the platform has successfully hosted nine modules across eight departments and over 1,000 students have access.
Lambda Feedback works by automating feedback on mathematical problems and solutions that students are set. Though currently utilising deterministic algorithms without using generative AI, Lambda Feedback has great opportunities to utilise generative AI in offering feedback on tasks and providing feedback on what to do next. In the early phase of using this platform, teachers receive the spread of answers that students have given and suggest feedback for the platform, meaning the Lamba Feedback tool can learn the ways in which to address correct and a spread of incorrect answers.
The research behind Lambda Feedback centres on our understanding of the capabilities of computers to undertake the tasks of teachers in university settings. The work has identified tasks such as finding errors in answers, writing feedback and recording progress, which could all be undertaken on a platform such as Lambda Feedback, giving time back to teachers. This means that in-person teaching time can be of the highest quality, with time only being used to discuss work and problems in a way that can only be undertaken by humans. The long-term goal of Lambda Feedback is to develop high-quality and highly personalised feedback at the time when homework is submitted, allowing students to receive constructive feedback instantly.
Find out more and see Lambda Feedback in action.
Sign Up
If you are interested in receiving the Learning and Teaching Newsletter, please email ltstrategy@ic.ac.uk.
You can also view the issue archive online.
