Imagine an athlete who is preparing to compete. In the lead up to the big event, they receive little or no coaching. But after crossing the finish line their coach gives them notes on how they could have improved their performance.
The flaws in this approach are obvious. Yet when it comes to ‘high performance writing’, this is the scenario university students face in the typical assessment cycle: it is only after a student submits an assignment that they receive feedback on how they could improve.
Since writing is time-consuming to read and comment on, the type of timely, personalised feedback athletes receive from coaches isn’t usually possible in the university context – but that could be about to change.
The UTS Connected Intelligence Centre (CIC) in partnership with Xerox Research Centre Europe (XRCE) are currently piloting a new software program that gives students automated feedback on their written assignments – at the draft stage. This means students can use the feedback to improve their content before submitting the assignment.
Called Academic Writing Analytics (AWA), the program provides feedback on students’ academic writing, highlighting areas where they’ve mastered skills like critical thinking and reflection.
“We’re using what’s called natural language processing, or NLP, which is the use of computers to read text and try and make sense of it,” says CIC Director, Professor Simon Buckingham Shum.
When a student enters a draft document into the AWA app, the program flags aspects that bear the hallmarks of good academic writing and suggests improvements for weaker content.
This initiative tackles the challenges facing students who enter university with undeveloped academic writing skills, or with a lack of experience using the language of the discipline they’re studying.
Academic Writing Analytics (AWA) program (screenshots by CIC)
Timely, actionable feedback is key to improving these capabilities, and it’s here that AWA is filling a much-needed gap. With some classes numbering in the hundreds of students, teaching staff don’t have the bandwidth to provide individual feedback on drafts. Academics from across UTS faculties and schools have partnered with the CIC team and literacy experts in UTS’s Institute for Interactive Media and Learning (IML) to familiarise the program with discipline-specific terminology, making it relevant to students across a broad range of course areas.
“This is a very multidisciplinary project, so we bring expertise in NLP and in designing user interfaces, but you also need the experts in reflective writing and how to teach it well, so we work with colleagues in IML,” Buckingham Shum says.
“It’s critical to work with the academics themselves – we involve them in the design process so that they can try the tool out and see how well it’s performing on the particular kind of writing they give their students, and then we tune the parser accordingly.”
AWA is the result of a partnership with Xerox Research Centre Europe that gives the CIC team access to XRCE’s NLP technologies in language understanding.
External collaboration and open innovation are at the heart of the way we work and do research.
This collaboration strengthens our international network far beyond Europe and demonstrates our ability to apply our natural language understanding methods to exciting real-life problems related to education.
— Dr Ágnes Sándor, project leader, XRCE
Research team
-
Director, Connected Intelligence Centre
