Education Research PhD Studentship in Trustworthy AI in education: large language models for automated dialogic feedback in STEM education

Applications are invited for a research studentship in the field of Trustworthy AI in Education leading to the award of a PhD degree. The post is supported by a bursary and fees (at the UK student rate) provided by Imperial College London. The successful stydent will be based in the Department of Mechanical Engineering

The project is to develop the use of Large Language Models (LLMs) to provide dialogic feedback, in particular when students use natural language to express ideas using domain-specific technical knowledge. For example, on the applicability of certain mathematical relations to a given problem, or discussions on climate change.

The research builds on developments in Education at Imperial College London providing automated feedback to students on STEM study. Our web platform already serves thousands of students with self-study content and automated feedback. The current PhD will extend our feedback algorithms beyond mathematical expressions and into natural language.

We want to explore the use of fine tuning to language models to embed the following into the behaviour of the model: educational principles of how to provide good dialogic feedback; domain-specific expertise; technical correctness.

The focus of the research, once the technical deployment begins, will be on student emotions as they interact with the system. We expect to deploy models very early in the process, and to refine their use based on feedback from users.

Candidates require advanced skills in computing, but also a strong motivation to contribute to progress in higher education. This will require critical thinking, an interest in deploying and evaluating solutions with users, and participating in educational research.

You will be an enthusiastic and self-motivated person who meets the academic requirements for enrolment for the PhD degree at Imperial College London. Candidates from any technical discipline are invited, on the condition that they have demonstrable skills in computing, deep technical expertise in a particular science/engineering discipline, strong intellectual capacity, and discipline in work habits. Teamwork and communication skills are essential.

To find out more about research in this area at Imperial College London in this area, go to: 
https://teachingengineers.wordpress.com/2022/07/18/computers-make-us-human/

To apply for the opportunity, please complete this form.