Machine learning and natural language processing in support of interactive automated tutoring for non-native

Massive online open courseware, adaptive learning management systems and the like are well-established, but their impact outside of a few subjects such as computer science and maths has been modest so far. The missing ingredient and essential difference between on-line maths or computer science and language learning courseware is objective automated and meaningful learning-oriented assessment (LOA).

LOA, to be useful, must include both a summative and formative component, must address the appropriate linguistic tasks and skills, must be interpretable and actionable by the learner, and must be integrated into an adaptive and therefore personalized learning platform.

I’ll describe how we have exploited (deep) machine learning and natural language processing to develop LOA for non-native writers, building accurate automated graders benchmarked to Cambridge English exams as well as pedagogically useful feedback models to support learners interactively and incrementally improve their writing.

Our writing assessment technology is already in use in Cambridge English courseware such as Empower and Linguaskill, and in Write and Improve (e.g: https://www.youtube.com/watch?v=LpTF_o_eyao).