Paper: Enhancing Peer Feedback in MarkEd

After a semester spent marking second-year students' essays, I proposed a dissertation project aimed at improving assessment practices in education.

Under the guidance of my amazing supervisors

Cristina Alexandru ↗ and

Aurora Constantin ↗, we extended MarkEd to address common issues in peer assessment. It combines structured moderation from human markers with AI-generated guidance to help students write more effective and reflective peer feedback.

The system was trialled in a large second-year course at the University of Edinburgh with 194 students. Of the students who received AI suggestions, 75% agreed or strongly agreed that they prompted reflection. The dissertation itself received a grade of Outstanding (93%).

That work became a paper, presented at the UK and Ireland Computing Education Research (UKICER) 2025 conference and published in the ACM conference proceedings.

UKICER 2025 · ACMRead the paper

From your friend,
Tomas