Explainable machine learning and natural language processing for automatic assessment

Topic Description

This project uses natural language processing and machine learning to develop methods for automatically marking students' answers to assessment questions.

Assessment is a key part of the education process, either for the learner to receive feedback on his or her progress, or for the educator to determine the progress of the learner. Automated testing often uses multiple choice questions, but there are many opportunities for assessment in which the learner responds in free text (such as a short sentence of English). Such questions are very difficult to mark automatically, because of the sophisticated linguistic analysis needed, or costly and error-prone to mark manually.

The (human) marker interacts with a machine learning system to collaboratively develop an assessment model. This has the twin advantages that the human marker retains ultimate control of how the marks are awarded, and that the grades awarded by the machine are more accurate than those based on machine learning systems alone.

Our current work uses inductive logic programming as the machine learning framework, although you would be free to use your own choice of machine learning technology if it is more appropriate.

Skills Required:

A strong background in either natural language processing or machine learning. Knowledge of the other area is desirable, but not essential.

Good programming skills.

Background Reading:

Willis, Alistair (2015). Using NLP to support scalable assessment of short free text responses. In: Proceedings of the Tenth Workshop on Innovative Use of NLP for Building Educational Applications, pp. 243–253. ( https://oro.open.ac.uk/43458/ )

Dzikovska, Myroslava O., Rodney D. Nielsen, and Claudia Leacock. "The joint student response analysis and recognizing textual entailment challenge: making sense of student responses in educational applications." Language Resources and Evaluation 50.1 (2016): 67-93. ( https://bit.ly/2ETx6x9 )

https://en.wikipedia.org/wiki/Inductive_logic_programming