by Dr Helen Yannakoudakis & Dr Ardeshir Geranpayeh
In recent years there has been a widespread change in the understanding of the relationship between learning and assessment. The perception of learning and assessment as being diametrically opposed is increasingly recognised as a false dichotomy. There is growing awareness of how assessment can be used to promote learning. This approach is encompassed in the concept of Learning Oriented Assessment (LOA).
LOA emphasises the primacy of the task cycle in any learning context: learning outcomes, language activity, observation, feedback and adjustment. The use of new technologies can be incorporated into this cycle to make it more effective and efficient. The use of connected devices allows tasks to take place outside of the classroom to open up the possibility of learning affordances happening anywhere at any time.
In this presentation we will demonstrate how by combining artificial intelligence with task-based pedagogy we have developed a teaching tool which uses automated writing assessment to provide diagnostic feedback to learners to enable and encourage them to improve their writing. We discuss the challenges of developing an automated placement model for writing. How does the artificial intelligence behind the model learn to mark as accurately as a trained examiner? What data does the artificial intelligence learn from? And, indeed, how do you create this artificial intelligence?
In the LOA approach the learner is at the heart of assessment and learning, and the tool has been designed and built to create different learning opportunities and effective learning outcomes. We will share results, of trails and feedback, which show how the tool has been used by learners to improve their writing.
Dr Helen Yannakoudakis, Senior Research Associate, Computer Laboratory, University of Cambridge, UK
Helen is a Senior Research Associate at the Computer Laboratory of the University of Cambridge, working on Automated Language Teaching and Assessment. She is also a fellow at Girton college, a committee member of Women@CL, and a Research & Development specialist at iLexIR. She holds a PhD in Natural Language and Information Processing from the University of Cambridge, during which she also worked on the English Profile Programme in collaboration with Cambridge English Language Assessment; an MPhil in Computer Speech, Text and Internet Technology; and a BSc in Computer Science. Her research interests lie at the intersection of Computational Linguistics, Machine Learning and Visualisation, and particularly revolve around the areas of automated language teaching and assessment, self-assessment and tutoring systems, error detection and correction, and second language acquisition.
Dr Ardeshir Geranpayeh, Head of Automated Assessment & Learning, Cambridge English Language Assessment, UK
Ardeshir Geranpayeh is the Head of Automated Assessment & Learning at Cambridge English Language Assessment, a department of the University of Cambridge. Ardeshir holds a PhD from the University of Edinburgh on the comparability of language proficiency testing. He has 27 years’ experience of test validation and has contributed to the design, development, revision and evaluation of several internationally recognised language proficiency tests. Ardeshir has published extensively on language proficiency testing and cheating detection.