Artificial intelligence offers solution to heavy marking loads

An assessment and feedback platform that can cut the time taken to mark coursework, homework and examination papers in maths and science subjects has been designed at the University of Birmingham.

Called Graide, the system is aimed at higher education institutions and has arisen from a PhD thesis written by Robert Stanyon and supervised by Professor Nicola Wilkin in the University’s School of Physics and Astronomy.

What sets Graide apart from existing computer aided assessment tools is its ability to learn an assessor’s marking style and it can assess not only final answers, but also a student’s workings. In addition, it accepts both handwritten and digital responses, does not require programming, and can fit into existing workflows. It is already being piloted at the University of Birmingham and the University of Liverpool.

Stanyon recognised the need for this sort of software during his time in higher education. Alongside co-developers, Manjinder Kainth and George Bartlett, Stanyon spent time working as post-graduate teaching assistant as part of his PhD studies. The insights into student needs they gained during this experience was fundamental to the development of Graide.

The three students have now co-founded a company called 6 Bit Education to commercialise the software.

“Assessment and feedback are a longstanding staple of teaching, but until recently, very little has changed about the basics of this process,” says Dr Stanyon, now CTO of 6 Bit Education. “Automated marking is an attractive option but existing Computer Aided Assessment tools generally require assessments to be modified to be compatible with the CAA platform. Also they are not able to mark elements such as mathematical working – they can only grade the final answer.”

The system developed by the team includes algorithms that learn an individual assessor’s marking style, so they do not have to grade the same answer twice. As the assessor progresses through marking an assignment, Graide will start to automate more and more of the feedback until all marking is complete.

The development team estimate that the system can reduce grading times by an average of 89 per cent. At the same, students receive more than 7 times as much feedback.

Manjinder Kainth, CEO of 6 Bit Education, said: “Providing high quality feedback is essential for students, but it is time-consuming to deliver and can be plagued by inconsistencies. Assessment and feedback is consistently one of the worst performing categories in the national student survey and this is the challenge that Graide is able to address.”

Professor Wilkin, also Director of Education in the University’s College of Physical and Engineering Sciences and Education Consultant to 6 Bit, said: “The experience of being a post-graduate teaching assistant, coupled with the students’ skills in maths and physics meant that this project was built on a genuine, well-understood need. This software has potential to shake up the way assessments in Maths and Science are carried out and I’m excited to see it take off.”

6 Bit Education has raised £800,000 from Mercia Asset Management and angel investors and has also embarked on a pilot scheme across an additional four universities with Jisc’s national centre for AI in tertiary education.